Abstract
Objective. Deficiency of adenosine deaminase 2 (DADA2) is an autosomal recessive autoinflammatory disorder associated with ADA2 mutations. We aimed to investigate the characteristics and ADA2 enzyme activities of patients with DADA2 compared to non-DADA2 patients.
Methods. This is a descriptive study of 24 patients with DADA2 who were admitted to the Adult and Pediatric Rheumatology, Pediatric Haematology, and Pediatric Immunology Departments of Hacettepe University. All ADA2 exons were screened by Sanger sequencing. Serum ADA2 enzyme activity was measured by modified spectrophotometric method.
Results. Twenty-four patients with DADA2 were included: 14 with polyarteritis nodosa (PAN)-like phenotype (Group 1); 9 with Diamond-Blackfan anemia (DBA)-like features, and 1 with immunodeficiency (Group 2). Fourteen PAN-like DADA2 patients did not have the typical thrombocytosis seen in classic PAN. Inflammatory attacks were evident only in Group 1 patients. Serum ADA2 activity was low in all patients with DADA2 except one, who was tested after hematopoietic stem cell transplantation. There was no significant difference in ADA2 activities between PAN-like and DBA-like patients. In DADA2 patients with one ADA2 mutation, serum ADA2 activities were as low as those of patients with homozygote DADA2. ADA2 activities were normal in non-DADA2 patients. ADA2 mutations were affecting the dimerization domain in Group 1 patients and the catalytic domain in Group 2 patients.
Conclusion. We suggest assessing ADA2 activity along with genetic analysis because there are patients with one ADA2 mutation and absent enzyme activity. Our data suggest a possible genotype–phenotype correlation in which dimerization domain mutations are associated with PAN-like phenotype, and catalytic domain mutations are associated with hematological manifestations.
Deficiency of adenosine deaminase 2 (ADA2) enzyme (DADA2) is an autosomal recessive autoinflammatory disease associated with mutations in ADA2 encoding the ectoenzyme ADA21,2. The ADA2 enzyme is essential for endothelial stability and a factor defining the macrophage subsets3. It is a dimeric protein secreted in the extracellular environment mostly expressed by monocytes and other cells of the myeloid lineage4,5. It has 4 domains: the signal sequence, the dimerization domain, the putative receptor-binding (PRB) domain, and the catalytic domain4. Dimerization of ADA2 monomers is required for full enzymatic activity and the dimer gets stabilized by interacting with the cell-surface glycosaminoglycans (GAG)6. The PRB domain is responsible for the binding of ADA2 to specific receptors while the catalytic domain is important for the catalytic function of the enzyme6.
The more common phenotype of DADA2 is vasculopathy and stroke7,8,9. The hematological and immunological findings of DADA2 include Diamond-Blackfan anemia (DBA)–like phenotype7,10,11, lymphopenia12, neutropenia9,11,13–17, pancytopenia10,11,15, Coombs positive hemolytic anemia10, autoimmune cytopenia2,9,11,16,18, hemophagocytic lymphohistiocytosis2,19, cellular immunodeficiency12, and hypogammaglobulinemia2,7,11,19.
Herein, we report the features of a cohort of patients with DADA2 from Hacettepe University, Ankara, Turkey. We also investigate the diagnostic potential of serum ADA2 activity measurement. In addition, we compared the laboratory test results of polyarteritis nodosa (PAN)–like DADA2 patients and non-DADA2 PAN patients.
MATERIALS AND METHODS
Patients
This is a descriptive study consisting of two DADA2 cohorts (n = 24) from Hacettepe University (Ankara, Turkey), which is a tertiary referral hospital in a country with high consanguinity rate. There is a strong collaboration between different departments of the medical school.
Group 1 included 14 DADA2 patients with the PAN-like phenotype. We have also included PAN patients (n = 60) who were negative for ADA2 mutations. We compared the results of basic laboratory tests [hemoglobin (Hb), white blood cell (WBC) count, thrombocyte count, erythrocyte sedimentation rate (ESR), C-reactive protein (CRP)] during active disease between PAN-like DADA2 patients and non-DADA2 PAN patients. For patients with DADA2, the response to etanercept (ETN) is defined as absence of any disease-related symptoms and normal acute-phase reactants (APR) without increase in doses of concomitant therapies such as corticosteroids.
Group 2 included 10 patients with DADA2, 9 diagnosed with DBA in their primary centers (the pediatric hematology departments from 5 different centers in the country) and referred for molecular workup and followup to our center. They were diagnosed with DADA2 and evaluated in our center at least once at the time of this study. These 9 DADA2 patients with DBA-like phenotype were among the 60 patients in our national DBA registry, all of whom had been tested for ADA2 mutations. One patient was diagnosed in the Department of Pediatric Immunology when she was screened for immunodeficiency genes in a panel.
Twelve patients with DADA2 had been included in previous papers2,19,20 and 2 other papers, including 2 patients from Group 2, have been submitted.
The serum ADA2 activity was also investigated in a group of patients who had features mimicking DADA2 (n = 89). This group included patients with systemic vasculitis other than PAN (e.g., Behçet disease), cytopenia (due to reasons other than DBA; e.g., myelodysplastic syndrome), immunodeficiency (e.g., CD4 deficiency), or a history of stroke or thrombosis (due to reasons other than DADA2; e.g., glial tumor). These patients were not tested for ADA2 mutations because their diagnoses were already confirmed. In addition to these patients, ADA2 activity was tested in 5 DBA and 3 PAN patients who did not have ADA2 mutations.
Written consent from the patients was obtained according to the Declaration of Helsinki. The study was approved by the ethics committee of Hacettepe University (GO 15/721).
Total ADA (tADA) and ADA2 activity assays
Two ml of venous blood were taken from patients, parents, and controls, and centrifuged at 1500 g for 15 min. Serums were separated and stored at −80°C until the analysis was performed.
Serum tADA and ADA2 activities were measured by making the total reaction volume half of the spectrophotometric method described by Giusti and Galanti21. This method is based on measuring the ammonia released during the conversion of adenosine to inosine, by the Berthelot reaction. A 21-mM adenosine solution in 50 mM phosphate buffer, pH 6.5 was used as substrate for tADA activity. Added to the test tubes was 0.1 mmol/l erythro-9-(2-hydroxy-3nonyl)adenine (EHNA), pH 6.5, an ADA1 inhibitor, in the presence of 21 mM adenosine for ADA2 activity22. A 15-mM ammonium sulfate standard stock solution was prepared in ammonia-free distilled water and then freshly diluted to 75 µM during the working day with 50 mM phosphate buffer, pH 6.5 and used as standard for the assay. Serum sample (25 µl) and 21 mM adenosine solution were mixed for tADA assay. For the ADA2 assay, 25 µl of serum sample, 21 mM adenosine, and 0.1 mM EHNA (at the final volume) were mixed. Then the reaction tubes were incubated at 37°C for 60 min in a water bath. After that, the reactions were stopped by phenol/nitroprusside (106 mM phenol; 0.17 mM sodium nitroprusside) solution. The serum samples were added to the blank tubes. Then alkaline hypochloride (11 mM NaOCl; 125 mM NaOH) solution was added to all test tubes, mixed, and incubated at 37°C for 30 min in a water bath. All samples were studied in double. The blue-violet color formed at the end of the experiment was measured at 628 nm using Shimadzu UV-1700 spectrophotometer. The results were calculated after subtracting sample blank values from sample values and reagent blank values from standard values. The results were expressed as U/l at 37°C. One unit of the tADA or ADA2 was defined as the amount of enzyme required to release 1 µmol of ammonia per min under standard measurement conditions. Percentage contributions of ADA2 activities to the tADA activity (ADA2%) were also calculated.
Genetic analysis
Fourteen PAN-like DADA2 patients (Group 1) were diagnosed with Sanger sequencing of 10 ADA2 exons at the National Institute of Health (n = 3) and Hacettepe University (n = 11). The same method was used for checking ADA2 mutations in the patients of the PAN cohort (n = 60). Primer sequences are presented in Supplementary Table 1 (available from the authors on request). PCR products were directly sequenced using ABI Prism 3130 Automated Sequencer (Applied Biosystems). MEFV gene variant analysis was also performed in these 14 patients with Sanger sequencing. Twelve variants [E148Q, P369S, F479I, M680I(G-C), M680I(G-A), I692del, M694V, M694I, K695R, V726A, A744S, R761H] were tested in MEFV in our Department of Medical Genetics.
In Group 2, one patient was diagnosed with whole exome sequencing and 8 were diagnosed with Sanger sequencing of ADA2 exons. One patient from the Department of Pediatric Immunology was diagnosed with an immunodeficiency gene panel analysis23. This panel included 266 genes that were analyzed with targeted next generation sequencing. Analysis was performed by Ion Reporter 5.6 Software.
Statistics
Statistical analysis was performed using the SPSS software (version 21.0; IBM Corp.). The variables were investigated using visual (histogram, probability plots) and analytic (Kolmogorov-Smirnov) methods to determine the normality of distribution. The data of descriptive analysis were expressed as the median, minimum, and maximum values. Categorical variables were compared with the chi-square test or Fisher’s exact test where appropriate. The Mann-Whitney U test was used to compare the non-normally distributed continuous data between 2 groups.
RESULTS
Demographic and clinical characteristics of patients with DADA2 (n = 24)
The median (min–max) ages at symptom onset and diagnosis were 3.5 years (1 mo–35 yrs) and 14 years (1.6–45 yrs), respectively. Male/female ratio was 1:1.
Group 1 included patients with PAN-like features (n = 14; Table 1 and Table 2). One also had severe anemia and pancytopenia. Group 2 included 9 patients with DBA-like features and 1 patient with immunodeficiency (Table 3).
There was parental consanguinity in 15 patients (62.5%). There was an affected sibling in 2 families.
Clinical and laboratory characteristics of patients in Group 1 (n = 14)
The main features of Group 1 patients are summarized in Table 1 and Table 2. All patients had recurrent episodes of fever, abdominal pain, and elevated APR. The APR were also increased in all during active disease.
Nine patients (64.2%), 5 adults and 4 children, fulfilled the Ankara 200824 and the American College of Rheumatology criteria25 for PAN.
There was proteinuria in 8 patients. It was in the nephrotic range in 3 [patients with focal segmental glomerulosclerosis (FSGS), amyloidosis, and mesangial proliferative glomerulonephritis], while mild proteinuria was present (∼180–200 mg/day) in 5. Proteinuria improved in all with ETN except 2, who died soon after DADA2 diagnosis.
In the PAN cohort without ADA2 mutations (n = 60; M/F = 1.4), there were 37 children (61.5%) and 23 adults (38.5%). The median (min–max) thrombocyte count of PAN patients was 440 (180–726) × 109/l, whereas it was 333 (129–502) × 109/l in PAN-like DADA2 patients (p = 0.009). There was no significant difference between these 2 groups in Hb, WBC count, ESR, and CRP levels.
Clinical and laboratory characteristics of patients in Group 2 (n = 10)
The clinical and laboratory characteristics of patients in Group 2 are summarized in Table 3.
The hemogram findings at diagnosis of pure red cell aplasia (PRCA) in patients 1–9 revealed a median Hb level of 4.3 g/dl (2.02–6) with a reticulocyte count of 0.2% (0.17–0.57%). Median WBC and platelet counts were 9.7 × 109/l (5.1–11.3) and 345 (204–690) × 109/l, respectively. PRCA was diagnosed with paucity of erythroid precursors in the bone marrow aspirate in an otherwise normocellular bone marrow in these patients. Parvovirus PCR was negative in all (peripheral blood).
Patient 10 had an afebrile convulsion during followup while she was taking ETN and still neutropenic. The magnetic resonance imaging was compatible with posterior reversible encephalopathy syndrome (PRES), while fungal abscess, encephalitis, or a vasculitic feature of DADA2 were excluded. ETN was discontinued because it has been implicated in PRES.
Serum tADA, ADA2, and ADA2% enzyme activities
Serum tADA and ADA2 activities were analyzed in 123 individuals: 16 DADA2 patients, 10 family members (parents/siblings) of DADA2 patients (all heterozygotes for ADA2 mutations), 3 PAN and 5 DBA patients who did not have ADA2 mutations, and 89 diseased controls (described above).
The median (min–max) ADA2 activities were 1.03 (0–2.6) in PAN-like DADA2 patients (n = 10); 0.91 (0.34–2.14) in DBA-like DADA2 patients (n = 5); 14.84 (13.05–27.13) U/l in PAN patients without ADA2 mutations (n = 3); 19.01 (15.39–39.19) U/l in DBA patients without ADA2 mutations (n = 5); 10.43 (7.95–15.07) U/l in heterozygote family members of DADA2 patients (n = 10); and 22.27 (13.08–57.35) in the diseased control group (n = 89). Thus, serum ADA2 activity was within normal range in PAN and DBA patients who did not have ADA2 mutations and other diseased controls (n = 89; Figure 1). There was no significant difference between PAN-like and DBA-like DADA2 patients regarding ADA2 activity (p = 1).
The ADA2% was 19.62 (0–75.29) in PAN-like DADA2 patients (n = 10) and 15.96 (5.86–20.61) in DBA-like DADA2 patients (n = 5). However, ADA2% was 72.66 (57.32–93.7) in heterozygote family members (n = 10), 78.88 (72.03–96.18) in DBA patients without ADA2 mutations (n = 5), 77.57 (75.36–85) in PAN patients without ADA2 mutations (n = 3), and 83.78 (37.76–100) in the diseased control group (n = 89). ADA2% was decreased in patients with DADA2.
ADA2 deficiency was identified in all patients with DADA2 but one. Patient 4 from Group 2 had ADA2 measurement after hematopoietic stem cell transplantation (HSCT) and ADA2 activity was normal.
In 3 patients (Group 1) who had responded to ETN, serum ADA2 activities remained very low subsequent to the treatment (range of pretreatment and posttreatment ADA2 activities 0.46–1.44 U/l and 0.46–1.43 U/l, respectively). At the time of the first evaluation of ADA2 activities, the disease was active in these patients. They had inactive disease under ETN for ≥ 6 months at the time of the second check of enzyme activities.
ADA2 activities of unaffected heterozygous parents were about half the activity of the control subjects, although these individuals were asymptomatic (Figure 1). However, in two DADA2 patients with one ADA2 mutation, ADA2 activities were as low as the activities of DADA2 patients with biallelic mutations.
Molecular workup
All DADA2 patients had ADA2 mutations. In Group 1, eleven were homozygous for G47R mutation in ADA2, two were heterozygous for G47R, and one was compound heterozygous for G47R and G47V (Table 4). These mutations affect the dimerization domain of ADA2 enzyme. MEFV variant analysis (in all Group 1 patients) revealed heterozygosity for E148Q in 2 patients.
The heterogeneous spectrum of ADA2 mutations in Group 2 patients were affecting the catalytic domain of ADA2 enzyme (Table 4).
Treatment: Group 1
All patients were refractory to corticosteroids. All responded to ETN except 2 who died soon after DADA2 diagnosis and 1 who is stable while taking colchicine.
Among pediatric patients, the patient with a previous diagnosis of cutaneous PAN (Patient 7) responded to mycophenolate mofetil and had been stable for 1 year; then her disease flared and ETN was started. Fresh frozen plasma (FFP) was tried in 2 siblings (Patients 8 and 9) while they were waiting for the approval of biologics; one did not respond while the other was stable while receiving FFP. Subsequently, ETN was started in both. All pediatric patients are currently stable while taking ETN.
Among the adult patients, 3 currently are stable while taking ETN. FFP was administered to 2; none responded. One adult patient has been clinically stable for the last 7 years while taking colchicine only; anti–tumor necrosis factor (TNF) treatment has now been suggested. Two patients died soon after DADA2 diagnosis; one before the initiation of ETN, the other 3 weeks after the initiation of ETN.
Treatment: Group 2
Four (44%) of 9 patients with DBA-like phenotype responded to corticosteroid treatment. HSCT was performed in 2 patients (one responsive, the other refractory to corticosteroid). Two patients after HSCT and another 2 after 2-year corticosteroid treatment achieved remission and required no transfusions thereafter. Four patients (refractory to corticosteroid) are currently in a transfusion program, along with iron chelation treatment with deferasirox. None of the 9 patients with DBA-like features has received ETN. In Patient 10, the neutropenia did not resolve with corticosteroids, cyclosporine, and immunoglobulin therapy. ETN offered temporary benefit with an increase in neutrophil count for a 2-week period, but it failed to resolve the neutropenia. Neutropenia did not respond to FFP, either. Matched unrelated donor transplantation was planned for severe persistent neutropenia because no matched family donor was found.
Outcome: Group 1
Eleven patients are clinically stable with normal CRP on ETN. All of these patients reached clinical and laboratory remission within the first month of ETN and they have not experienced any disease flare so far while using ETN. One patient is stable while taking colchicine only with normal APR (checked every 3 mos). Two adult patients (Patients 1 and 2) died soon after DADA2 diagnosis.
The disease course in Patient 1 was complicated with myelofibrosis and secondary amyloidosis (renal and intestinal). He was resistant to the immunosuppressive treatments and FFP, and died of necrotizing pneumonia 2 months after DADA2 diagnosis. Tuberculosis was ruled out. The family did not consent to autopsy.
Patient 2 did not respond to various immunosuppressive therapies and FFP, and developed resistant digital ulcers and FSGS (collapsing variant) while he was taking immunosuppressives and iloprost. He died despite the initiation of ETN and plasmapheresis.
Outcome: Group 2
Four patients are in remission while taking no treatment (2 after HSCT; 2 after 2-year corticosteroid treatment). One patient who responded to corticosteroids is still receiving corticosteroid treatment. No relapse has been observed in the aforementioned 5 patients. Four patients who did not respond to corticosteroid treatment are currently in a transfusion program. Patient 10 still has severe neutropenia refractory to treatment.
DISCUSSION
To our knowledge, this is the first study reporting the whole DADA2 spectrum with PAN-like, hematological, neurological, and immunological presentations.
The definition of DADA2 changed the diagnosis for a significant number of patients who had been originally diagnosed with classic PAN. Further, PAN-like features with recurrent fever and abdominal pain attacks in DADA2 might have suggested the co-occurrence of familial Mediterranean fever (FMF) and PAN. Thus, our patients with PAN-like DADA2 had been tested for MEFV mutations and the results had not confirmed an FMF diagnosis. Two of them were heterozygous for E148Q, the frequency of which is around 6% in a healthy Turkish population26,27.
We have suggested that low to normal thrombocyte count in DADA2 may help in differentiating from PAN where thrombocytosis is expected. Thrombocytosis in classic PAN generally occurs as a reactive change due to inflammation28. The absence of thrombocytosis in PAN-like DADA2 is interesting. It also remains unknown how DADA2 causes DBA-like anemia. ADA2 may have a role in hematopoietic process besides its inflammation-associated effects.
In our patients with DBA-like DADA2, the corticosteroid responsiveness was 44%, while it is around 80% in classic DBA patients in the literature29. On the other hand, hepatosplenomegaly, which was present in 4 out of 9 patients with DBA-like DADA2, is not a feature of DBA, which is a bone marrow failure syndrome29. Additionally, compared to classic DBA29, dysmorphic features were more subtle in DADA2 and immunologic abnormalities were common.
Proteinuria was observed in more than half of patients with PAN-like DADA2, mostly as mild asymptomatic proteinuria in the absence of a renal disease. This may be caused by endothelial dysfunction or chronic inflammation30. Its improvement with ETN suggests a predominant role for inflammation in the pathogenesis of proteinuria.
The catalytic domain of ADA2 was affected in DBA-like patients while the dimerization domain was affected in PAN-like patients in our cohort. Mutation in the dimerization domain can impair the dimer formation and interaction with GAG, leading to decreased ADA2 activity. However, catalytic domain mutations probably cause conformational change and impair the catalytic function of the enzyme6. ADA2 plays a role in inflammation by promoting macrophage differentiation, inducing proliferation of monocytes, and inducing the expression of genes in neutrophils coding for some proinflammatory cytokines4. However, its role in the pathogenesis of a DBA-like phenotype remains largely unknown. Based on the site of the mutation, the functions of ADA2 may be affected at different levels. It is important to note that there are reports of DADA2 patients with catalytic domain mutations who presented with PAN-like phenotype in the absence of immunologic and hematologic findings2,31. Thus, further analyses are necessary for the clarification of the genotype–phenotype correlation.
We had 2 patients with DADA2 who had one ADA2 mutation; heterozygotes with the phenotype have previously been reported32. Possible explanations are the presence of a mutation in the promotor region or other interacting genetic factors. However, the fact that the enzyme activity was at an in-between level in the healthy heterozygotes while it was way below normal in the DADA2 patients with one ADA2 mutation indicates the advantage of enzyme testing. We now suggest that a test be done first for the enzyme activity in siblings of patients with DADA2.
Enzyme levels are crucial in the initial stage of diagnosis; however, there is no practical use of enzyme levels for followup of patients with DADA2. Persistently low ADA2 activities in patients with DADA2 who are in clinical and laboratory remission with anti-TNF treatment suggests that the clinical features of these patients are seen through inflammatory pathways, possibly related to TNF. Caorsi, et al have already shown higher TNF secretion from stimulated monocytes of patients with DADA2 compared to healthy controls32. Anti-TNF treatment probably suppresses the inflammation induced by the inflammatory macrophages in DADA2 but is not expected to normalize the ADA2 activity, which was shown in our study as well. Our results also suggest that discontinuing anti-TNF treatment may not be reasonable unless ADA2 activity is restored with a curative therapy such as HSCT.
We suggest that remission in patients with DADA2 could be described as no clinical features and normal APR. It is not certain whether we can taper the dose of anti-TNF therapy in remitting patients. Lifelong anti-TNF treatment is indeed challenging. Further, it is not suitable for the patients with DBA-like phenotype. HSCT seems a curative therapeutic option for DADA2, especially in patients with predominant hematologic and immunologic features33. In a cohort of 14 patients with DADA2, HSCT was performed for bone marrow dysfunction and/or immunodeficiency and was shown to restore ADA2 enzyme activity11. Larger cohort studies are required to reach a firm conclusion about the curative potential of HSCT in DADA2.
There are several unmet needs in DADA2, such as the need for prenatal diagnosis or testing asymptomatic siblings. We confirm that enzyme activity can guide us in deciding on therapy in patients in whom only one ADA2 mutation has been shown.
There are a few limitations of our study. DADA2 patients with one ADA2 mutation might have another mutation in ADA2 that we were not able to detect. In addition, it is not possible to totally exclude DADA2 diagnosis without genetic testing in the diseased controls with normal ADA2 activity.
This study highlights the diagnostic potential of ADA2 activity measurement in suspected patients. We also suggest a phenotype–genotype correlation in our cohort. Further, we suggest for the first time that a simple laboratory test such as platelet count may help for differentiating DADA2 from PAN.
Acknowledgment
We are grateful to European Rare Diseases Research Program, which supported DBA molecular investigations as part of the EuroDBA Project. Professor Nurten Akarsu performed the molecular studies of DBA patients; special thanks to her for her great support in EuroDBA. We thank Prof. Zeynep Karakas from Istanbul University; Associate Professor Ikbal Ok from University of Health Sciences, Ankara Child Health and Diseases Hematology Oncology Training and Research Hospital; Associate Professor Muge Gokce from Kanuni Sultan Suleyman Hospital; and Dr. Ali Fettah from Sami Ulus Children’s Hospital, who sent their DBA patients’ data and blood samples for molecular investigation.
Footnotes
Supported by Scientific and Technological Research Council of Turkey (TÜBITAK) with grant number 315S192.
- Accepted for publication April 16, 2019.