Introduction

Chronic granulomatous disease (CGD) is a rare (1:200,000–1:250,000 births in USA) inherited disorder caused by defects in the nicotinamide adenine dinucleotide phosphate oxidase (NADPH) complex [1–3]. The NADPH oxidase complex is composed of two membrane proteins (gp91phox and p22phox), three cytosolic components (p47phox, p67phox, and p40phox), and the Rac GTPase [4].

CGD affected patients suffered frequently from recurrent, life-threatening bacterial and fungal infections, with 2/3 of cases presenting their first manifestations at an early age, sometimes even at birth [5, 6]. Pneumonia is the most common form of infections, although abscesses and lymphnoditis are also frequently observed [4, 6–8]. In North America, infections are mainly arised from five pathogens, specifically, Staphylococcus aureus, Serratia marcescens, Burkholderia cepacia, Nocardia, and Aspergillus [1, 5]. Mycobacterium tuberculosis and Bacille Calmette-Guerin (BCG) infections are frequent in other areas of the world [2, 3, 7, 9], such as Turkey and Hong Kong. Apart from infections, CGD patients are vulnerable to granulomatous formation and autoimmune diseases [1, 10, 11].

To date, CYBB (encoding gp91phox), NCF1 (encoding p47phox), CYBA (encoding p22phox), NCF2 (encoding p67phox), and NCF4 (encoding p40phox) genetic mutations have been reported in CGD patients [4, 12]. Approximately 65 %–70 % of patients worldwide have X-linked CGD (X-CGD) caused by CYBB (Xp21.1) mutations, but in countries with a higher incidence of consanguineous marriage, autosomal recessive CGD (AR-CGD) is the most common form [1, 4]. Neutrophils functional tests (either nitroblue tetrazolium test (NBT) or the dihydrorhodamine (DHR) assay) are rapid tools for the diagnosis of CGD [13, 14], although the DHR assay is more sensitive and preferable [1]. Genetic analysis can identify the disease-causing genes in patients, find carriers and enable development of prenatal tests. Effective prenatal diagnosis (including direct sequencing of DNA from amniotic fluid cells or chorionic villus samples (CVS) and DHR assays on umbilical cord blood) is an important tool for controlling the number of children born with CGD [15, 16].

Conventional CGD treatments include prophylactic antimicrobial agents, surgical drainage or excision [6], and administering interferon-γ (IFN-γ). Hematopoietic stem cell transplantation (HSCT) is the only cure for CGD. Despite improvements in treatment, CGD still has a high mortality rate (2 %–5 % per year in USA) [17].

To date, there have been no reports in a large sample of Chinese children with CGD. Thus, in this study, we described the clinical features, laboratory tests, genetic analysis, prognosis and outcome of 38 CGD patients from mainland China, and perform prenatal diagnosis in eight fetuses.

Methods

Patients

During 2003 to 2013, we enrolled 38 cases (including 3 transplanted patients) diagnosed with CGD in the Children’s Hospital of Chongqing Medical University. This study was approved by the ethics committee of Children’s Hospital of Chongqing Medical University and informed consent obtained from the children or their parents.

Basic Information and Laboratory Tests

Diagnoses of tuberculosis (TB) included clinically diagnosed and confirmed cases (Diagnostic criteria for pulmonary tuberculosis, China (WS 288–2008)). Clinical diagnosis was made according to clinical manifestations, radiography, laboratory tests (PPD, ESR, TB-Ab, T-SPOT.TB, and lymph node biopsy) and effective anti-tuberculosis treatments. Diagnosis of confirmed cases involved detecting the presence of Mycobacterium tuberculosis by either sputum smear microscopy, sputum, pus or gastric liquid cultures, TB-DNA-PCR, or lung biopsy. Diagnosis of BCG disease (regional, extra regional localized, disseminated disease, and other BCG syndromes) was made as described by Hesseling AC et al. [18]. A “probable” diagnosis of disseminated BCG disease was made as described by Lee PP et al. [9].

Methods used to distinguish between TB and BCG vaccinations included spoligotyping and multiplex PCR [19].

DHR Flow Cytometry Assay

DHR flow cytometry assays were performed as previously described [15]. Data were analyzed using CellQuest software (BD FACSCaliburTM Flow Cytometer).

DNA and RNA Extraction

Genomic DNA (gDNA) and total RNA from patients and family members were isolated from whole blood cells according to the manufacturer’s recommendations. Total RNA was reverse transcribed into cDNA using the PrimeScriptTM RT reagent Kit with gDNA Eraser (Perfect Real Time) (TaKaRa, Dalian, China).

PCR, RT-PCR and Sequencing

Relevant genes were amplified from gDNA and cDNA using specific oligonucleotide primers (Sangon Biotech, Shanghai, China). PCR products were confirmed by gel electrophoresis and then sent to Sangon Biotech (Shanghai) Co., Ltd for sequencing.

Sequence Analysis

Sequence analyses were performed using Bioedit software and the basic local alignment search tool (BLAST) from the National Center for Biotechnology Information (NCBI; http://blast.ncbi.nlm.nih.gov/Blast.cgi). Before identifying novel missense mutations, 100 normal alleles were sequenced to rule out the possibility of single nucleotide polymorphisms (SNPs).

Prenatal Diagnosis

Female carriers were advised to have prenatal diagnosis performed when they were pregnant again. Amniotic fluid (30 ml) was collected at 18 weeks gestation using ultrasound guidance. DNA and RNA were extracted from amniotic fluid cells. Direct sequencing of relevant gDNA and cDNA regions were performed. Simultaneously, the sex determining region Y (SRY) gene was PCR amplified to determine the gender of the fetus. If direct sequencing failed to determine if the fetus was normal or not, umbilical cord blood (2 ml) was collected at 26 weeks gestation to detect the respiratory burst of neutrophils by DHR assay. After birth, DHR assays were performed to verify the accuracy of the prenatal diagnosis.

Results

Patient Demographics

We enrolled 38 CGD patients (36 male, 2 female) from 36 unrelated families (patients 11 and 34 are brothers and patients 17 and 18 are twin brothers), and 14 provinces and cities of mainland China (Sichuan 29 %, Chongqing 21 %). The type of onset in this cohort included recurrent fever (68.4 %; 26/38), cough (28.9 %; 11/38), and abscess (18.4 %; 7/38). The mean age at onset was 3.4 months (range 2 days–18 months), and the mean age of diagnosis was 31.7 months (range 18 days–172 momths). To date, 17 patients (44.7 %) have died, with the mean age of death being 30.9 months (range 6 months–102 months). Thus, 20 patients are alive (including 2 transplanted patients), with a mean age of 71.1 months (range 3 months–206 months). However, patient 1 did not participate in the follow-up period of the study. Approximately 78.9 % (30/38) of the CGD patients were diagnosed between 2009 and 2013.

Overall, 39.5 % (15/38) had positive family histories including known family histories of CGD and early death due to recurrent infections. The mean diagnosis age of patients with positive family histories was 32.3 months (range 0.7 months–172 months), similar to those with negative family histories (31.3 months, range 1.9 months–116 months).

Infection

In this cohort, 32 cases (84.2 %) developed their first clinical manifestations before 6 months. The most common infections were pneumonia (84.2 %; 32/38) and abscesses (84.2 %; 32/38) including skin, perianal, liver, and spleen abscesses. Additionally, 36.8 % (14/38) suffered form recurrent diarrhea, but only 3 children presented with colitis confirmed by colonoscopy, 44.7 % (17/38) with hepatosplenomegaly, and 42.1 % (16/38) with lymphadenitis. The main pathogens were Staphylococcus aureus and Candida albicans.

The most common types of Candida albicans infections were thrush (7 cases) and mycotic pneumonia (5 cases). We identified Candida albicans in six patients (15.8 %; 6/38). Results of fungal cultures with sputum, bronchoalveolar lavage fluid, or feces showed Candida albicans in 4 cases, 1 case, or 1 case, respectively.

BCG Complications

In total, 32 children received BCG vaccination (BCG strain D2 PB302) at birth or later as part of the National Vaccination Program. Fifteen cases did not suffer from BCG disease, while seventeen cases had BCG disease including regional (10 cases), extra regional localized (17 cases), disseminated (4 cases), and probable disseminated (10 cases). 21 patients presented with pulmonary tuberculosis including 19 clinically diagnosed and 2 confirmed cases. The common site of occurrence for disseminated BCG disease was the lungs.

Laboratory Features

Routine blood tests, immunoglobulin levels, NBT, and DHR flow cytometry assays were shown in Table 1. Overall, 36 patients had elevated white blood cells (WBC), mainly neutrophils. All the children except three cases had elevated C-reactive protein (CRP). Blood sedimentation tests were performed on 26 patients, and 21 had high erythrocyte sedimentation rates (ESR). Considering their age, 32 patients had remarkably high serum IgG levels, and some had high serum IgE levels. The NBT (0 %–3 %) and neutrophil oxidative function were all significantly lower than health controls.

Table 1 Laboratory features of 38 CGD patients in mainland China
Full size table

Gene Analysis

In total, we tested 36 male children for CYBB gene mutations, and identified 29 different mutations in 31 patients, including 14 reported mutations and 15 novel mutations (51.72 %). There were 12 splice errors (including 4 suspected splicing errors) (patient 31, Fig. 1), 7 deletion mutations, 5 nonsense mutations, 4 missense mutations and 1 large gene deletion (XK, CYBB and DYNLT3 genes) . Two splicing products were identified in cDNA from patient 8 (CYBB c.483-484ins115bp, p.K161fsX13; and c. 338-483del146bp, 483-484ins115bp, p.A113fsX3), while his cousin was diagnosed with CGD in the fetal period by prenatal diagnosis but had different splicing products (CYBB c.483-484ins115bp, p.K161fsX13; and c.483-484ins115bp, 484-674del191bp, p.K161fsX13). A GT homozygous deletion in NCF1 (c.75-76 del GT) was found in Patient 27 and 32, and a heterozygous NCF1 mutations (c.764-801del 38 bp and 923C > T) in patient 30. In the CYBA gene, 1 homozygous missense mutation and 1 homozygous deletion mutation were found in patients 26 and 33, respectively. No CYBB gene mutations were identified in gDNA, and because patient 1 did not participate in the follow-up period, the cDNA of CYBB gene and other genes were not detected. Although CYBB, CYBA, NCF1, NCF2 and NCF4 were all analyzed, no mutations were identified in patient 20, and additional novel genes may be required for further analysis of this patient. Overall, we identified 31 carriers in 23 unrelated families (Table 2).

Fig. 1
figure 1

CYBB gene mutation analysis of patient 31 and his mother. (a). A nucleotide substitution (IVS6 + 1333 T > G) located far from any splice sites, leading to insertion of a section of intron 6 (80 bp) between exons 6 and 7 (b). Mother was a heterozygote at this point (IVS6 + 1333 T > G) and cDNA sequence was normal

Full size image
Table 2 Genetic mutations and carriers of 38 CGD patients and their families in mainland China
Full size table

Prenatal Diagnosis

Combined with direct sequencing analysis, gender identification and DHR assays, we identified 4 fetuses with CGD (from families 8, 11, 19 and 26) and 4 healthy fetuses (from families 6, 8 and 19).

Treatment and Outcome

Upon diagnosis, all CGD patients were advised to receive prophylactic treatment with cotrimoxazole (24 mg/kg/d, divided twice daily) and itraconazole (patients age < 13 years or < 50 kg, 100 mg/d; > 13 years or > 50 kg, 200 mg/d) to prevent bacterial and fungal infections.

By the end of December 2013, three patients had undergone HSCT. The first patient (patient 18) was given an HLA-B and HLA-CW mismatched unrelated cord blood (CB) graft after receiving busulfan (BU) + cyclophosphamide (CY) + anti-thymocyte albumin (ATG) as a pretreatment regimen. Unfortunately patient 18 had acute graft-versus-host disease (GVHD, grade two) affecting the skin and gut, followed by four bouts of pneumonia at 3 months (Respiratory Syncytial Virus and Cytomegalovirus), 6 months (fungi and Respiratory Syncytial Virus), 10 (Moraxella catarrhalis), and 22 months after transplantation. He suffered from infectious diarrhea twice at 3 months (Enterobacter aerogenes) and 15 months (Pseudomonas aeruginosa) after transplantation. 13 months after transplantation, the child developed septic arthritis (Pseudomonas aeruginosa), and 22 months after transplantation died of severe pneumonia, acute respiratory distress syndrome, and sepsis (Acinetobacter baumannii, fungi, Cytomegalovirus, and Epstein Barr virus). After a conditioning regimen with BU + CY, the second patient (patient 20) was transplanted using a HLA-identical sibling donor, and is now healthy and doing well. The last patient (patient 28) was also transplanted using a HLA-matched sibling donor. He had grade-1 GVHD and presented with sepsis after transplantation. DHR assay were recovered normal in the first two children, but presented with a mosaic pattern in the third one after transplantation.

Discussion

Here we described detailed clinical features, laboratory tests, gene analysis and outcome of 38 CGD patients from mainland China. In our cohort, children suffered from first infections at an early age (between 0 and 6 months), but diagnosis was delayed significantly (between 13 and 40 months). Lack of awareness and screening tests for CGD may be important reasons for this. With significant progress in medicine in China, the diagnostic rate is higher than before, with approximately 78.9 % of CGD patients were diagnosed in the past 5 years.

In accordance with previous studies, pneumonia was the most common form of infections [3, 20, 21]. We also frequently detected abscesses, mainly skin abscesses (42.1 %). This differs from the USA [21] and Israel [22], where deep abscesses, including liver, spleen, brain or pulmonary abscesses, are the major type (46 %–76 %).

A significant feature of our cohort was a high incidence of mycobacteria infections. Based on our findings and those in previous studies [3, 9], it appears that CGD patients are prone to TB and BCG complications. The types of BCG disease include regional, extra regional localized, disseminated disease, and other BCG syndromes [18]. Disseminated BCG disease is the most serious complication, and associated with a high mortality rates (60 %–83 % in immunocompromised patients) [23]. We found 55.3 % and 10.5 % of cases suffered from pulmonary TB and disseminated BCG disease, respectively. In addition, we did not identify BCG in 10 cases (26.3 %), therefore these patients likely have probable disseminated BCG disease. Similarly, 55.9 % of CGD patients in Iran presented with BCGitis [20]. Moreover, in a study from Hong Kong, 41.2 % (7/17) and 47.1 % (8/17) patients presented with TB and BCG infections [9]. The high incidence of BCG infections demonstrates that the respiratory burst of neutrophils plays an important role in limiting BCG replication. Apart from CGD, BCG infections are reported in other primary immunodeficiency diseases (PID), including severe combined immunodeficiency syndrome (SCID), Mendelian susceptibility to mycobacterial diseases (MSMD), and X-linked hyper-IgM syndrome (XHIM) [24]. The pathway of IL-12/23-IFN-γ plays a crucial role in the host response to infections with mycobacteria. SCID and MSMD caused by deficiency in these pathways, so the patient with SCID and MSMD often suffered from the most critical complications of BCG. In CGD, BCG lymphadenitis is the most common form of BCG complications. CGD patients are usually respond well to anti-TB regimens (excluding pyrazinamide as BCG is uniformly resistant to it) [25] compared with SCID and MSMD. In our study, the children with BCG disease were recommended to receive anti-TB treatments (isoniazid, rifampicin and ethambutol), but because of side effects (hepatorenal damage) and for financial reasons, most patients did not persist with regular treatment, and seven cases (50 %; 7/14) died of probable/disseminated BCG disease. Thus, we suggest that newborns with a family history of PID should avoid BCG vaccination. Patients with BCG infections are suggested to take immune function screening tests, including the respiratory burst assay. In addition to mycobacterial infections, we also found Candida albicans infections were more common compared with other reports.

Hypergammaglobulinemia is a common finding in CGD patients [26, 27]. In our cohort, five cases had normal IgG levels, while the others had elevated IgG levels. Some cases had high serum IgE levels as described by Patiroglu T et a l [27].

There were 31 patients (31/38; 81.6 %) with X-CGD, but because of the higher incidence of consanguineous marriage in other parts of the world (e.g. Middle Eastern countries, North Africa, and much of western, central, and southern Asia), AR-CGD is the most common form. In a study of 38 CGD patients form Israel [22], Baruch Wolach et al. demonstrated autosomal recessive inheritance in 63 % of cases. Similarly, in Iran, 87.1 % of cases inherited a form of AR-CGD [20].

In a previous study, 20 %–30 % of all cases resulted from defects in NCF1 (A47°CGD), the most common form of AR-CGD [28]. In contrast to other CGD types, there is a “hot mutation” (a GT homozygous deletion, △GT) at the beginning of exon 2 in NCF1. We detected this mutation in patients 27 and 32. Recombination between ψNCF-1(NCF-1 pseudogenes) and NCF1 functional gene causes high incidence of the △GT mutation. To date, NCF1 has at least two pseudogenes, both of which show highly homology to NCF1 and coexist on chromosome 7q11.23 [28, 29]. A remarkable feature of ψNCF-1 is the GT deletion at the beginning of exon 2, as if this deletion occurs in NCF1 it can cause CGD. In addition to this, there are other difference between NCF1 and ψNCF-1 as Paul G. Heyworth et al. described [28].

We identified four suspected splicing errors in our cohort, but no mutations in coding regions or splice sites of the CYBB gene, in gDNA for patients 8, 16, 22, and 24. However we identified a large fragment deletion in CYBB cDNA, indicating that mutations may occur in intronic regions outside the splice sites. In patient 31 (Fig. 1), we found a nucleotide substitution (IVS6 + 1333 T > G) located far from any splice sites, leading to insertion of a section of intron 6 (80 bp) between exon 6 and 7. This is an unexpected splicing pattern, similar as described by Noack D et al. [30]. Thus in genetic analysis, detection of mutations in cDNA is better than that of gDNA for the discovery of splicing errors.

Patient 8 and his cousin had X-CGD caused by different mutations in the CYBB gene. We speculate that a de novo mutation occurred during embryogenesis in the maternal grandmother, with unsuccessful DNA repair resulting in two aberrant alleles. Therefore, the maternal grandmother of patient 8 had triple somatic mosaicism (a normal allele, the mother’s allele, and the aunt’s allele). One of the two aberrant alleles has been inherited by patient 8 and his mother, and the other by his cousin and aunt. Unfortunately at present, we have not identified the carriers (including the mother, aunt, and maternal grandmother), but we will perform further studies in this family with this intention [30, 31].

Regarding treatment for CGD, prophylaxes with antibacterial and antifungal agent are used worldwide, although the efficacy of IFN-γ therapy is still in dispute [11, 32]. Research by the International Chronic Granulomatous Disease Cooperative Study Group found that IFN-γ treatment was effective regardless of the inheritance mode, and if taking antibiotics or not [33]. However, a long term follow-up of 60 CGD patients by Martire et al. [11], demonstrated that IFN-γ treatment did not reduce the incidence of infections (severe and mild infections), and that there are adverse events (e.g. fever, myalgia, and headache) to adherence to IFN-γ prophylaxis. In our cohort, cotrimoxazole and itraconazole were recommended and IFN-γ was not routinely used. Seventeen cases (44.7 %) died of various infections at an early age due to diagnosis delay, unclear pathogens, and substandard treatment. Until recently, Aspergillus species infections were the main cause of mortality in CGD patients [17]. We found only one patient had an Aspergillus infection, although it is perhaps due to limitations in the detection methods that we did not detect more evidence for Aspergillus infection in the other cases. There is controversy about optimum timing and indications in HSCT. As previously discussed, transplantation with an HLA-matched sibling donor at an earlier age is the best choice [17, 32]. However, in a multicenter study by Tayfun Güngör et al. [34], 56 CGD patients (75 % patients with high-risk features) were enrolled. With reduced-intensity conditioning regimens, all 21 HLA-matched related-donor and 35 HLA-matched unrelated-donor (HLA-9/10 or HLA-10/10 match with the recipient) transplants had a better outcome (overall survival rate 93 %, event-free survival rate 89 %), suggesting a reduced-intensity conditioning regimen is safe and effective for high-risk CGD patients with HLA-matched related- or unrelated-donors.

Prenatal diagnosis is an effective way to control the birth of children with CGD. Direct sequencing of gDNA/RNA from amniotic fluid cells or CVS can provide an earlier diagnosis, but careful attention must be paid to avoid contamination with maternal DNA. Although DHR-assays are another reliable and accurate method for prenatal diagnosis, umbilical cord blood cannot be collected until 18 weeks gestation [16]. Whereas a prior study described a lower than normal respiratory burst of neutrophils at 22 weeks gestation, in our study, DHR-assays were performed at 26 weeks gestation (using a male fetus from family 8 and a female fetus from family 26). A combination of direct sequencing and DHR assays is an effective and accurate way for prenatal screening for CGD.

There are some shortcomings in our study, specifically, this is a single center study with the patients mainly from Sichuan and Chongqing (50 %), and therefore may not completely and accurately describe CGD in China. In this regards, we are setting up a system of “CARE CENTER FOR PID IN CHINA”. In our cohort, most patients had no etiological evidence of CGD; this may be due to limitations in detection methods.

Conclusion

We have described the clinical and molecular features of 38 CGD patients from mainland China. A remarkable feature of our cohort is a high incidence of TB and BCG infections. Molecular analysis is an important tool for identifying patients, carriers and high-risk CGD fetus.