Chronic granulomatous disease: Clinical, microbial, and genetic findings in 39 Colombian patients

Chronic granulomatous disease (CGD) is an inborn error of immunity (IEI) characterized by abolished or impaired activity of the nicotinamide adenine dinucleotide phosphate (NADPH) oxidase complex, which is essential for the generation of reactive oxygen species (ROS) and microbial killing. The clinical manifestations of CGD typically appear during infancy or early childhood (1). Patients with CGD are particularly susceptible to Staphylococcus aureus, Burkholderia cepacia, Escherichia coli, Nocardia spp., Pseudomonas spp., Serratia marcescens, Salmonella spp., and Klebsiella spp. Infectious diseases due to Mycobacterium tuberculosis, the mycobacterium responsible for tuberculosis, or Mycobacterium bovis–Bacille Calmette-Guérin (BCG) vaccine have also been documented in CGD patients. The organs most frequently affected are the lungs, lymph nodes, liver, bones, gastrointestinal tract, and skin (2, 3). Fungal infections are also frequent and mostly caused by filamentous molds. Invasive aspergillosis due to Aspergillus fumigatus is frequently documented, followed by infections due to Aspergillus nidulans, Aspergillus terreus, and Aspergillus tanneri (4, 5). In addition to this susceptibility to infection, CGD patients display a predisposition to the development of granuloma and dysregulated chronic inflammation (1, 4, 5, 6, 7, 8, 9). CGD can be diagnosed biologically through assays evaluating NADPH oxidase function, including the nitroblue tetrazolium test, ferricytochrome-c reduction, and various chemiluminescent or fluorescent probe–based methods. The dihydrorhodamine 123 (DHR) assay, a fluorescence-based test quantifying hydrogen peroxide production and performed by flow cytometry, has become the preferred method in routine clinical laboratory diagnosis (10).

Genetically, CGD results from loss-of-function or rare hypomorphic variants of any of the genes encoding components of the phagocyte NADPH oxidase complex. This multicomponent enzymatic complex includes three cytosolic subunits—p47^phox, p67^phox, and p40^phox—encoded by NCF1, NCF2, and NCF4, respectively; two membrane-bound components—gp91^phox and p22^phox—encoded by CYBB and CYBA, respectively; and the chaperone protein EROS (encoded by CYBC1/C17ORF62), which is required for the stability and assembly of the gp91^phox–p22^phox heterodimer (8, 9, 10, 11, 12, 13, 14, 15, 16). Rare hemizygous variants of CYBB are the most common genetic cause of CGD, underlying the X-linked recessive form of CGD (XL-CGD) (OMIM #306400), which accounts for approximately two thirds of cases worldwide (3, 8, 17, 18, 19, 20, 21). As a result, 70–80% of affected individuals are male. The remaining autosomal recessive (AR) forms are caused by rare homozygous or compound heterozygous variants of CYBA (OMIM #608508), NCF1 (OMIM #608512), NCF2 (OMIM #608515), NCF4 (OMIM #613960), and CYBC1 (OMIM #618334) (1, 8, 10, 11, 12, 13, 22, 23, 24, 25). The incidence of CGD has been estimated at 1 per 200,000–250,000 live births and varies between countries worldwide (19, 20, 26, 27). However, the true incidence of CGD in Colombia remains unclear due to the absence of large, systematically characterized cohorts. Several major contributions from Latin America, including efforts to establish registries in multiple countries, have expanded regional knowledge of the disease, but specific data for Colombia remain limited and are derived primarily from isolated case reports and small registry-based series (3, 18, 28, 29, 30). This study addressed this gap in our knowledge, through characterization of the clinical, microbiological, and genetic features of Colombian patients with CGD.

We investigated 39 individuals from 32 unrelated Colombian kindreds enrolled in this study (Table 1 and Fig. 1 a). The patients originated from 12 different departments of Colombia: Cundinamarca (n = 13), Antioquia (n = 6), Nariño (n = 3), Valle del Cauca (n = 2), Bolívar (n = 2), Tolima (n = 2), and Cauca, Huila, Arauca, Córdoba, Boyacá, and Santander (n = 1 each). The regional origin of five patients was not specified (Fig. 1 a). 35 patients (89.74%) were male and four (10.25%) were female. Consanguinity was documented in two patients (P8 and P19), and one patient belonged to an endogamic community (P10); all three individuals had AR-CGD (Table 1). The median age at symptom onset was 3 months old (mo) (interquartile range [IQR] 1–8.5), and median age at CGD diagnosis was 23 mo (IQR 6.5–39). Whole-exome sequencing (WES) or Sanger sequencing was performed on the probands whenever feasible (Table 1). Eight patients did not undergo genetic testing and were diagnosed on the basis of clinical manifestations and a DHR assay performed several years earlier (Table 1). 22 patients carried hemizygous CYBB variants, including hemizygous missense (n = 3), nonsense (n = 11), essential splice site (n = 4), and large deletion variants (n = 4). Most of these variants have already been reported, but three of the CYBB were new: c.563T>G predicted p.L188*, c.935T>A predicted p.M312K, and c.1314 + 1_+5del. We also identified one female patient (P32) with a CYBB defect displaying skewed X inactivation (c.1226C>T; p.A409E). Three patients (7.7%) were homozygous (essential splice site) or compound heterozygous (small deletion and essential splice site) for pathogenic variants of CYBA, including the previously unreported c.59-1G>A variant. One patient was homozygous for the very common pathogenic variant c.75_76del (often written c.75_76delGT) of NCF1, consistent with AR-CGD. One patient was homozygous for a previously reported pathogenic variant of NCF2 (c.233G>A; p.G78Q), associated with AR-CGD. Finally, one patient (P14) was homozygous for a variant of NCF4 also carried by three asymptomatic siblings (P12, P13, and P15) in the homozygous state (Table 1) (9). These findings highlight the predominance of XL-CGD in this Colombian cohort, the broad geographic distribution of affected families, and the presence of both classic AR forms and a separate subgroup carrying NCF4 variants.

The DHR assay, which measures phorbol 12-myristate 13-acetate (PMA)–induced ROS production by neutrophils, is the standard functional test for the diagnosis of classic CGD (9, 34). The percentage DHR oxidation was determined for all patients (Fig. 1 b, data not shown in 20 patients) and was zero or markedly lower than that in healthy controls in 35 CGD patients; this assay also revealed the distinctive profile of p40^phox deficiency, with DHR activity in response to PMA stimulation remaining normal in four related individuals from kindred I (P12, P13, P14, and P15) (9). Impaired activation was observed when neutrophils from P14 and P12 were primed with low concentrations of tumor necrosis factor (TNF) and then stimulated with N-formyl-Met-Leu-Phe (fMLP) (data not shown). A more detailed assessment of intracellular superoxide production in P14 revealed a marked decrease following stimulation with serum-opsonized heat-killed Candida albicans (Fig. 1 c), heat-killed Histoplasma capsulatum (Fig. 1 d), or serum-opsonized zymosan (Fig. 1 e), whereas PMA-induced superoxide production remained relatively normal (Fig. 1 f). Together, these findings confirm that all Colombian patients with variants of CYBB, CYBA, NCF1, and NCF2 display impaired neutrophil NADPH oxidase–dependent responses, consistent with a diagnosis of CGD, except for those with the NCF4 variant, who displayed impaired ROS production in response to physiological stimuli.

Bacterial infections were the most common manifestations. Salmonella spp. isolation was reported for 17 patients (43.58%), with bacteremia documented in 12 of these patients (70.58%). Single cases of pulmonary nodules, brain abscess, lymphadenitis, and septic arthritis and three cases of sepsis were also observed (Table 2). Klebsiella spp., a member of the Enterobacteriaceae family like Salmonella, was identified in six infection cases (bacteremia, osteomyelitis, and pneumonia) (35). S. aureus was isolated from 15 patients (38.46%; Table 1). Other Gram-negative bacteria were also found (Fig. 2), including Pseudomonas spp., Pseudomonas fluorescens, Pseudomonas putida, B. cepacia, S. marcescens, Acinetobacter baumannii, and Enterobacter cloacae. We also observed several patients infected with another Gram-positive coccus, such as Enterococcus spp., or Leuconostoc pseudomesenteroides. Bacterial infections of the lungs (pneumonia, n = 29; 74.35%) were the most frequent, followed by bacteremia (n = 27; 69.23%), abscesses (n = 19; 48.72%), lymphadenitis (n = 18; 46.15%), and osteomyelitis (n = 8; 20.51%; Table 1). Clinical manifestations were compatible with CGD in 36 of the 39 patients described in this manuscript, with the other three individuals with NFC4 variants remaining asymptomatic to date (P12, P13, and P15). All patients were vaccinated with a single dose of M. bovis–BCG (strain Pasteur 1173), administered to the left arm, during the neonatal period or within the first month of life. Eight patients (P1, P10, P16, P22, P23, P30, P31, and P35; 17.95%) developed axillary lymphadenitis following vaccination (BCG-itis), progressing to disseminated BCG disease (BCG-osis) in one patient (P10). Pulmonary infection with M. tuberculosis was identified in four patients (P28, P30, P32, and P33) among these patients, only P30 also developed BCG-itis.

Nine patients (P8, P16, P18, P24, P25, P30, P32, P33, and P39; 23.08%) developed aspergillosis, and two others presented other fungal infections (Candida tropicalis, n = 1; H. capsulatum, n = 1). Two patients had documented viral infections (respiratory syncytial virus and parainfluenza) with moderate clinical manifestations, and one patient was infected with M. pneumoniae (Fig. 2 and Table 1). Granulomas were observed in eight patients (20.51%), inflammatory bowel disease in six patients (15.38%), autoimmune cytopenia in three patients (7.69%), and a macrophage activation syndrome secondary to infection was documented in one patient (2.56%; Table 1). Patients P16, P30, and P31 also presented progressive muscle weakness, consistent with the large deletion on the X chromosome also encompassing the DMD gene responsible for Duchenne muscular dystrophy. Together, these findings highlight the substantial infectious and inflammatory burden of CGD in these Colombian patients, while underscoring the clinical heterogeneity associated with different genetic defects.

All but four (p40^phox deficiency) of the patients received prophylaxis with trimethoprim-sulfamethoxazole and antifungal drugs, including itraconazole (n = 21, 53.84%), voriconazole (n = 13, 33.3%), or posaconazole (n = 3, 7.69%). 16 patients (41.03%) received human recombinant interferon-γ (hrIFN-γ) treatment in addition to antimicrobial therapy. 17 patients underwent hematopoietic stem cell transplantation (HSCT, 43.6%). Three patients (P16, P26, and P33) required a second HSCT, for a total of 20 procedures (n = 20). Median age at HSCT was 28 mo (IQR 11–74).

Two patients are currently awaiting HSCT. The donor was haploidentical, with posttransplantation cyclophosphamide in 11 transplants (55%) and umbilical cord blood (UCB) in eight transplants (40%), and one patient underwent HSCT from a matched sibling donor (5%) (Table 3). Myeloablative conditioning was used in seven transplants and reduced-intensity conditioning in 13 (65%). 16 of the transplant recipients received a busulfan and fludarabine conditioning regimen, associated with cyclophosphamide in three cases and melphalan in two (Table 3). Three transplant recipients received a busulfan-cyclophosphamide conditioning regimen, and four underwent total body irradiation. All patients received antithymocyte globulin for in vivo T cell depletion to reduce the incidence of graft-versus-host disease (GVHD) and graft failure. Graft failure occurred in six transplants (30%): five receiving cells from a haploidentical donor and one from UCB. Acute (a)-GVHD developed in eight transplant recipients: a-GVHD I (n = 1), a-GVHD II (n = 2), a-GVHD III (n = 2), and a-GVHD IV (n = 3). Chronic (c)-GVHD occurred in seven transplant recipients: c-GVHD I (n = 2), c-GVHD II (n = 3), and c-GVHD III (n = 2). One patient (P9) developed a thyroid tumor, and one patient (P24) presented posttransplant lymphoproliferative disease. Overall, these data reflect the heterogeneous therapeutic approaches used in this cohort, including broad antimicrobial prophylaxis and hrIFN-γ, and underscore the central but challenging role of HSCT, with the use of various conditioning strategies, frequent GVHD, and marked graft failure rates, in the management of CGD patients.

In total, 19 patients diagnosed with CGD died, including 12 patients with XL-CGD and one patient with AR-CGD; in six patients, the genetic defect was not identified. Fig. 3 presents overall survival (Fig. 3 a), differential survival for patients with XL-CGD versus AR-CGD (Fig. 3 b), and survival for patients who underwent HSCT compared to those who did not (Fig. 3 c). The leading cause of death was infectious complications, which occurred in 12 patients who had not undergone HSCT and seven patients who had undergone HSCT. The causes of death in the HSCT group included posttransplant infections (n = 2), combined GVHD and infection (n = 2), severe a-GVHD (n = 1), multiorgan failure (n = 1), and posttransplant lymphoproliferative disease (n = 1). All individuals with AR p40^phox deficiency (one symptomatic and three asymptomatic) remain alive (Table 1). In the 36 affected patients, excluding the three asymptomatic individuals with NCF4 variant, overall survival was higher in patients who underwent HSCT (n = 10 of 17 HSCT, 58.8%) than in those who did not (n = 6 of 18 non-HSCT, 33.3%). Median survival was similar for AR-CGD and XL-CGD patients (log-rank chi-square 2.822, P = 0.093). Hazard ratio was higher for XL-CGD 4.857 (95% confidence interval [CI] 1.46–16.11). Mortality was higher in patients with XL-CGD (12 out of 23, 52%) than in AR-CGD (1 out of 6, 17%), but it did not reach statistical significance (Fisher’s exact test P = 0.1834).

Median survival was similar for patients who underwent HSCT compared to those that did not (log rank chi-square 0.577, P = 0.448). Hazard ratio was 0.685 (95% CI 0.269–1.747). Mortality was higher in patients who did not undergo HSCT (12 out of 18, 67%) than in those who did (6 out of 16, 37.5%), but it did not reach statistical significance (chi-square 2.892, P = 0.0890). These findings highlight the high mortality burden associated with CGD in this cohort, together with the potential survival benefit of HSCT despite its inherent risks.

We report the first comprehensive cohort of Colombian patients with CGD. The geographic distribution of patients across 12 departments likely reflects referral patterns rather than true national prevalence. Geographically remote areas, particularly along the Andes mountain range and in southeastern jungle regions such as the Amazon and Orinoquía, are probably underrepresented due to barriers in healthcare access, limited specialist availability, and long travel distances. Genetic analyses demonstrated that XL-CGD was the most prevalent form, consistent with previous reports for Latin-American, Mexican, and U.S. cohorts, but contrasting with reports from Egypt, China, Morocco, and Turkey, where AR-CGD predominates (18, 23, 28, 29, 30, 36, 37). Notably, we identified variants not yet described in CYBB, specifically c.563T>G (predicted p.L188*), c.935T>A (predicted p.M312K), and c.1314 + 1_+5del, expanding the spectrum of known CGD-causing variants. We also identified a female patient (P32) heterozygous for the p.A409E variant with clinical CGD, highly impaired ROS production, and skewed X-chromosome inactivation on whole blood. Three patients with gp91^phox deficiency displayed a combined CGD–Duchenne muscular dystrophy (DMD) syndrome; in one case, the diagnosis of DMD was established only after HSCT. This dual phenotype is biologically and clinically plausible because CYBB and DMD are located in close proximity on Xp21.1, where large deletions can disrupt both genes simultaneously (38). Nine of the 39 patients (23%) had AR-CGD. Among them, three individuals carrying NCF4 variants are asymptomatic to date; therefore, these three individuals were excluded from the clinical and survival analyses. Other Latin American studies have found AR-CGD in 13–33% of cases, but in countries with a higher incidence of consanguineous marriages, this form can account for up to 60% of CGD cases (18, 23, 27, 30, 36, 39, 40, 41, 42, 43). The frequency of p47^phox deficiency in our cohort was 3.2% (1/39), which is lower than the 24.7% (71/287) reported by Kuhns et al. in North American and European populations (44) but closer to the 7.6% (18/238) described by de Oliveira et al. in a more comparable regional context (18). Although p47^phox deficiency is often associated with a milder clinical course that could potentially delay diagnosis, our cohort represents one patient referred to a tertiary immunology center. Furthermore, given the modest sample size, proportional comparisons should be interpreted with caution. Our genetic studies identified one previously unreported CYBA variant, c.59-1G>A. Two patients carried the c.177del CYBA variant; one was homozygous (P10) and the other was compound heterozygous (P9) with the c.59-1G>A/c.177del variant. These two patients came from neighboring departments in Colombia, suggesting a possible local founder effect. Furthermore, we detected no pathogenic variants of CYBC1, consistent with its relative rarity in CGD cohorts worldwide (6, 45, 46, 47). P14 presented with severe infections and was diagnosed with NCF4-associated CGD. His three siblings, aged 3, 8, and 10 years at the time of his diagnosis, were all found to be homozygous for the same NCF4 variant but have remained asymptomatic, demonstrating the variability of clinical expression of the disease, at least during adolescence. Continued longitudinal follow-up of the asymptomatic siblings will be essential to evaluate potential late-onset symptoms.

Median age at immunological diagnosis in our cohort was 23 mo, with an early median age at symptom onset, the main clinical manifestation being pneumonia, as reported in other studies (17, 18, 23, 27, 30, 36, 39, 40, 41, 42, 43). This probably reflects the predominance of XL-CGD, which is typically detected earlier in life and with more severe manifestations, together with ongoing improvements in the recognition of IEIs in Colombia. These improvements are being driven by increasing collaboration between clinical immunologists, infectious disease specialists, and pediatricians and by improvements in access to DHR testing and molecular diagnosis across the country. In our series, diseases due to Salmonella spp. were the most frequent infectious manifestation, consistent with findings from other Latin American, Mexican, and U.S. cohorts, but contrasting with the lower frequency of this infection reported in China, India, Turkey, Israel, Morocco, and Egypt (17, 18, 19, 21, 23, 27, 30, 36, 39, 40, 41, 42, 43). Salmonella spp. infection accounted for three deaths, highlighting the importance of this bacterium as a major clinical pathogen in patients with CGD in Colombia. Unfortunately, specific serotyping into typhoidal or non-typhoidal strains is not routinely performed in our clinical setting, limiting our ability to classify this severe infection. Klebsiella spp. infections were also frequent, consistent with the close phylogenetic and pathogenic similarities between Klebsiella and Salmonella (48, 49). Together, these findings highlight the need for early diagnosis, antimicrobial prophylaxis, and greater clinical suspicion of CGD, particularly in patients presenting with severe or recurrent Salmonella infections. One patient (P29) developed infection with L. pseudomesenteroides, an opportunistic microorganism mainly reported in hospitalized patients exposed to broad-spectrum antibiotics or invasive devices (50). No specific association with neutrophil defects has been established; therefore, this episode is more appropriately interpreted as potentially healthcare associated. All patients were vaccinated with BCG at birth, in line with Colombia’s universal immunization policy. Several patients were vaccinated despite a known family history of CGD, highlighting the need for the genetic counseling and explicit education of families concerning BCG-related risks. Confirmed or suspected M. bovis–BCG disease was observed in 17.95% of patients, a rate similar to that reported in Mexico, where this disease was observed in 30% of patients, but lower than that in the broader Latin American cohort (45%) (18, 30). Aspergillus spp. was identified as another important pathogen in our cohort, consistent with findings for other cohorts (18, 39, 45, 46, 47). Immune dysregulation was also observed, most notably inflammatory bowel disease in XL-CGD, consistent with the known inflammatory complications associated with this genotype (51). Overall mortality in our cohort exceeded 50%, but HSCT was associated with a marked improvement in survival (58.8 vs. 33.3%), confirming the curative benefits of this treatment, when available. The frequency of HSCT was higher in our series than in the most recently reported Latin American series (41 vs. 22%), although graft failure remained frequent in both series. In our cohort, the predominance of haploidentical and UCB transplants due to limited donor availability likely explains the increased graft failure incidence, rather than a population-specific biological effect. Reported graft failure rates after HSCT for CGD range from ∼5–20%, related to haploidentical or UCB grafts, particularly when reduced-intensity conditioning is used, low CD34⁺ cell dose, or significant pretransplant inflammation (52, 53, 54). In contrast, severe combined immunodeficiency generally demonstrates lower graft failure rates in matched donor settings, likely due to profound host immune deficiency facilitating engraftment (55, 56, 57). In our cohort, we observed marked clinical trends suggesting a worse prognosis for patients with the XL-CGD variant and for those who did not receive a HSCT. Although mortality was notably higher in patients without HSCT (67 vs. 37.5% in transplanted patients), this difference did not reach statistical significance (P = 0.089), likely due to the small sample size. However, it is important to highlight that the hazard ratio analysis did demonstrate that patients with XL-CGD have a significantly higher risk of mortality, almost five times higher, compared to the AR-CGD variant (Hazard Ratio 4.857; 95% CI 1.46–16.11). Altogether, these findings underscore the severity of the XL genotype and support the clinical benefit of HSCT, although multicenter studies with larger cohorts are required to statistically confirm the impact of transplantation on survival.

Expanding early access to immunological and genetic diagnosis, particularly in remote regions of Colombia, is essential to ensure timely referral for HSCT and to reduce preventable morbidity and mortality (18). Importantly, earlier diagnosis also increases the opportunities for genetic counseling, empowering families by helping them to understand inheritance patterns, reproductive risks, and preventive strategies, such as the avoidance of BCG vaccination in newborns at high risk. This study provides the first comprehensive characterization of CGD in Colombia, delivering crucial epidemiological and molecular insights that will enhance national diagnostic capacity and support better care for patients and their families across Colombia.

The patients and their relatives were living in Colombia (South America) and were followed up locally. They were recruited by the Group of IEIs (formerly the Primary Immunodeficiencies Group) in Medellin and at collaborating centers in Colombia. Written informed consent was obtained from the patients, their relatives, and the volunteer healthy controls enrolled in the study. Healthy volunteers were recruited at the University of Antioquia. This study was conducted in accordance with the “Scientific Standards for Technical and Administrative Health Research” established by Colombian Ministry of Health resolution 008430 of 1993. It was approved by the local review board of the University of Antioquia (F8790-07-0010) and by HOMI Fundación Hospital Pediátrico La Misericordia (735-24R). Experiments on samples from human subjects were performed in France and Colombia, in accordance with local regulations and with the approval of the corresponding institutional review boards.

Genomic DNA (gDNA) was extracted with the Puregene DNA Purification Kit (Gentra Systems) from whole-blood samples collected from patients and probands. Exome capture was performed with Agilent SureSelect V4/V5/V6 kits (Agilent Technologies), and paired-end sequencing was performed on a HiSeq 4000 platform (Illumina) generating 100-base reads. The reads were aligned with the reference human genome GRCh38/hg38 with the Burrows-Wheeler Alignment tool (BWA v. 0.7.12-r1039) (58). PCR duplicates were removed with Picard tools (https://picard.sourceforge.net/). The GATK base quality score recalibrator was applied to correct sequencing artifacts. GATK HaplotypeCaller was used to identify variant calls (59, 60). In some cases, depending on the patient’s health insurance cover, genetic sequencing was performed by external laboratories. Variants were confirmed by Sanger sequencing on gDNA from patients, parents, and relatives, when possible. Amplicons were sequenced with BigDye Terminator technology and a 3500XL Genetic Analyzer Sequencer (Applied Biosystems). We used the software analyzer Genius R9 v9.1.3 (Snapgene, GSL Biotech LLC) and 4Peaks (Nucleobytes B.V.) to align the sequences obtained with the reference sequence deposited in the NCBI database.

Whole blood from patients, unaffected relatives, and healthy controls was collected into heparin-containing tubes. Red blood cells were lysed with ammonium chloride solution. After washing, DHR (Sigma-Aldrich) and catalase (1,300 IU/ml; Sigma-Aldrich) were added, and the mixture was incubated at 37°C for 5 min. We then added buffered solution without stimulant or with PMA (400 ng/ml) (Sigma-Aldrich) to the tubes, which were incubated in a water bath at 37°C for 14 and 20 min, respectively. Samples were analyzed with a FACS SORT flow cytometer. We analyzed a total of 20,000 cells in the neutrophil gate (based on forward and side scatter) for fluorescence intensity in the FL-2 channel (61).

Luminol-enhanced chemiluminescence assays were performed on polymorphonuclear neutrophils (PMNs) to assess their NADPH oxidase activity. PMNs (2 × 10⁵ cells) were resuspended in PBS supplemented with glucose (PBS plus 0.9 mM CaCl₂, 0.5 mM MgCl₂, and 20 mM dextrose) and luminol (50 µM) and plated in 96-well plates in the presence or absence of superoxide dismutase (SOD; 75 µg/ml). Cells were incubated at 37°C for 10 min before the addition of PMA (400 ng/ml), opsonized heat-killed C. albicans, or 1 × 10⁶ heat-killed H. capsulatum cells. The number of relative light units (RLUs) was determined every 60–90 s over a period of 30 or 50 min with the long kinetic module in a Lmax microplate luminometer (Molecular Devices). Integrated RLU values were calculated with SoftMax software (Molecular Devices), with the subtraction of background chemiluminescence (measured in wells including all reagents but no cells) to obtain values for total ROS production.

Statistical analyses were performed with IBM SPSS Statistics for Windows, Version 25.0 (IBM Corp.), and survival data were analyzed using the Kaplan–Meier method using GraphPad Prism version 11. Given the sample size, qualitative data are presented as frequencies and percentages, whereas quantitative data are reported as medians and IQR. Categorical variables were analyzed with chi-squared tests. P values below 0.05 were considered statistically significant.

Informed consent for participation in this study was obtained in accordance with local regulations, with approval from the Institutional Review Board (IRB). The experiments described here were performed in Colombia and in France, in accordance with local regulations and with the approval of the IRB of Necker Hospital for Sick Children, France, and the local review board of the Universidad de Antioquia (F8790-07-0010).

Written informed consent to participate was obtained from all the patients.

Consent to publish this report was obtained from patients. All the authors approved the final version of the manuscript.

The data underlying this study are not publicly available due to patient privacy issues. The data are available from the corresponding author upon request.

We thank the patients, their families, and the physicians for their participation in the study. We are grateful to Deisy López, Manuela Tejada, Camila Carmona, and the Group of Inborn Errors of Immunity from Colombia for their contributions. We also thank Mónica Lucía Giraldo for her guidance on bioethical matters in Colombia and Yelena Nemirovskaya, Candace Clift, Dana Liu, Maya Chrabieh, Lazaro Lorenzo, and all members of the Laboratory of Human Genetics of Infectious Diseases (HGID) for their administrative and technical support.

The HGID laboratory is supported by the Howard Hughes Medical Institute, the Rockefeller University, the St. Giles Foundation, the National Institutes of Health (NIH) (R01AI127564, U19AI162568, R01AI095983), the National Center for Advancing Translational Sciences, NIH Clinical and Translational Science Award program (UL1TR001866), the French National Research Agency (ANR) under the “Investments for the Future” program (ANR-10-IAHU-01), the Integrative Biology of Emerging Infectious Diseases Laboratory of Excellence (ANR-10-LABX-62-IBEID), the French Foundation for Medical Research (EQU201903007798), Institut National de la Santé et de la Recherche Médicale, and Paris Cité University. This work was supported by the Colombian Administrative Department of Science, Technology and Innovation (code-111574455633, contract number 544-2016), ECOS-NORD (grant #63407-2018-C19S01-63407), Comité para el desarrollo de la investigación University of Antioquia—2017-16003, Ministerio de Ciencia, Tecnología e Innovación Colombia 111584467551, CT 415-2020, an Education European Society for Immunodeficiencies (ESID) 2018 Medium-term-fellowship, and the Fundación “Diana-García de Olarte” para las Inmunodeficiencias Primarias, Medellin, Colombia. J. Rojas is supported by the “Medellin Investiga” 2024 prize and an Education ESID 2025 Medium-term-fellowship.

Author contributions: Julian Rojas: data curation, formal analysis, investigation, methodology, visualization, and writing—original draft, review, and editing. Carlos A. Arango-Franco: conceptualization, data curation, investigation, methodology, project administration, validation, and writing—original draft, review, and editing. Marcela Moncada-Velez: investigation. Diana Marcela Arboleda: investigation. Edgar Alfonso Figueredo: data curation and formal analysis. Manuela Molina: conceptualization, formal analysis, and investigation. Juan Pablo Sánchez: investigation. Lina Marcela Zapata: data curation, methodology, resources, software, and validation. Paula Catalina Diez: investigation and methodology. Jesús Armando Álvarez: investigation. Maria S. Serna-Arbelaez: investigation. Lizet Jazmin Pérez Zapata: resources and writing—review and editing. Gabriel Vélez: investigation. Camille Soudée: investigation. Juan F. Alzate: methodology and software. Felipe Cabarcas: formal analysis. Catalina Obando Gil: investigation, resources, and writing—original draft, review, and editing. Carlos Garcés: conceptualization, data curation, and writing—review and editing. Martha I. Alvarez-Olmos: conceptualization, investigation, validation, and writing—review and editing. Paola Marcela Pérez-Camacho: writing—original draft, review, and editing. Diego Medina-Valencia: conceptualization, data curation, and writing—original draft. Jaime Alberto Patiño-Niño: resources. Juan Pablo Rendón: resources. María Claudia Ortega-López: investigation and resources. Luz Elena Cano: conceptualization, data curation, methodology, and writing—original draft, review, and editing. Juan Francisco López: conceptualization, investigation, supervision, and writing—review and editing. Flor Marcela Estupiñán: resources and visualization. Mauricio Chaparro: data curation and writing—review and editing. Laura Niño: resources. Andrés Felipe Escobar: conceptualization, investigation, resources, validation, and writing—review and editing. Jean-Laurent Casanova: formal analysis, funding acquisition, resources, and writing—review and editing. Stéphanie Boisson-Dupuis: resources and writing—review and editing. Anne Puel: resources, supervision, and writing—review and editing. Carlos Olmos: writing—review and editing. Luis Miguel Sosa: investigation and writing—review and editing. Patrick Eliana Sarmiento-Wilche: writing—review and editing. César Muñoz: investigation, resources, and writing—review and editing. Julio César Orrego: conceptualization, data curation, and visualization. Juan Álvaro López: conceptualization, investigation, methodology, resources, supervision, and visualization. Manuela Olaya-Hernández: conceptualization and writing—review and editing. Natalia Builes: resources, supervision, and writing—review and editing. Estefanía Vásquez-Echeverri: conceptualization, data curation, investigation, visualization, and writing—original draft, review, and editing. Natalia Vélez-tirado: supervision and writing—original draft, review, and editing. José Luis Franco: conceptualization, funding acquisition, investigation, project administration, resources, supervision, and writing—review and editing. Lina María Castaño-Jaramillo: conceptualization, data curation, formal analysis, resources, validation, visualization, and writing—original draft, review, and editing. Jacinta Bustamante: conceptualization, investigation, methodology, project administration, resources, supervision, validation, visualization, and writing—original draft, review, and editing. Andrés Augusto Arias: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, supervision, validation, visualization, and writing—original draft, review, and editing.