We evaluated a patient with a history of multiple episodes of recurrent meningitis caused by Neisseria spp. infections. The clinical presentation raised suspicion of an underlying inborn error of the complement system, prompting comprehensive immunological and genetic evaluation. In addition to the index case, close family members—including the patient’s siblings and mother—were included to investigate inheritance patterns and identify potential genetic variants associated with complement deficiency ( Table 1 ). The patient’s father was not included in the study, as he had passed away from colorectal cancer in 2021 at the age of 58. The family was of Caucasian ancestry, and the parents were not consanguineous.

Peripheral blood samples were obtained from all available participants. Total genomic DNA was extracted from peripheral blood mononuclear cells (PBMCs) for genetic analysis, while serum samples were used for functional assessment of the complement system.

This study involving human participants was reviewed and approved by the Portal de Ética de la Investigación Biomédica, Junta de Andalucía (Protocol Code: 0297-N-21). Written informed consent was obtained from all participants before inclusion.

Genomic DNA was analyzed using next-generation sequencing (NGS) with Twist Bioscience kits for molecular amplification and library preparation. Sequencing was performed on the Illumina NextSeq 1000 platform, and data analysis and variant interpretation were conducted using SOPHiA DDM™ for Genomics. Sequence alignment was carried out against the GRCh37/hg19 human reference genome. All identified variants of interest were subsequently validated by targeted Sanger sequencing.

The customized gene panels included the coding regions and flanking intronic sequences of 575 genes associated with primary immunodeficiencies, as previously described (18). These panels also encompassed all complement-related genes, including: C1QA C1QB, C1QC, C1QTNF4, C1R, C1S, C2, C2orf69, C3, C4A C4B, C5, C6, C7, C8A, C8B, C8G, C9. The complete gene panel used was included as Supplementary Materials ( Supplementary Table 1 ).

The diagnosis of meningitis was established through analysis of cerebrospinal fluid (CSF) and blood cultures. Samples were processed immediately upon receipt by the Department of Microbiology.

Blood cultures were performed using the BD BACTEC FX system (Becton Dickinson), which detected anaerobic bacterial growth within 16 hours. Gram staining revealed the presence of Gram-negative diplococci. Subsequent isolation was carried out on blood agar and chocolate agar, with incubation at 37°C in a 5% CO₂ atmosphere. After 24 hours, colonies were observed that were smooth, transparent, non-pigmented, and non-hemolytic on blood agar. The oxidase test was positive, and definitive bacterial identification was performed using MALDI-TOF mass spectrometry (MALDI Biotyper, Bruker).

The CSF sample appeared turbid, and Gram-negative diplococci were also observed on Gram stain. Cultures were processed in the same manner as the blood cultures, using both blood agar and chocolate agar, with incubation and identification following the same protocol.

Allele frequencies of the identified gene variants were obtained from publicly available population databases, specifically the Genome Aggregation Database (gnomAD) (19) and the Allele Frequency Aggregator (ALFA) database (20).

Serum C5 levels were measured by nephelometry using specific reagents from TRIMERO Diagnostics (Barcelona, Spain) on a BN™ II analyzer (Siemens Healthineers). The method is based on light scattering generated by immune complexes formed between C5 and anti-C5 antibodies. Results were expressed in mg/dL, using a standard calibration curve and internal quality controls provided by the manufacturer.

Total complement activity was assessed using the Autokit CH50 (Wako Chemicals GmbH, Germany), an automated, liposome-based immunoassay that measures activation of the classical complement pathway. The assay relies on the lysis of antibody-coated liposomes containing glucose-6-phosphate dehydrogenase (G6PDH). Upon complement-mediated lysis, G6PDH is released and reacts with NAD and glucose-6-phosphate (G6P), producing NADH, which is quantified by the increase in absorbance at 340 nm. Reactions were conducted at 37°C on an automated analyzer using 10 µL of patient serum 250 µL of reagent R1 (liposomes with G6PDH) and 125 µL reagent R2 (substrate with anti-DNP antibodies, NAD, and G6P). CH50 activity was automatically calculated and expressed in CH50 U/mL using a calibration curve. Internal quality controls were included in each analytical run. Serum samples were either processed immediately or stored at −70°C until analysis. All reagents were stored at 2–10°C according to the manufacturer’s instructions.

Genomic data was analyzed using the NCBI reference transcript NM_001735.3. The potential impact of the identified variants on protein function was assessed through a comprehensive set of in silico predictive tools, including PolyPhen-2 (21), Mutation T@ster2 (22), Mutation Taster 2021 (23), SIFT (24), LRT (25), MutationsAssessor (26), FATHMM (27), REVEL (28), MetaRNN (29), BayesDel addAF (30). The results of these algorithms are summarized in Supplementary Tables 2, 3 . Variant interpretation was further guided by the American College of Medical Genetics and Genomics (ACMG) classification criteria (31).

To explore the structural impact of the variants, in silico molecular modeling and structural analysis of the C5 protein were performed. A three-dimensional representation was used to assess how the amino acid substitutions could affect local structural environments and functional domains. PyMOL (version 2.5.7, Schrödinger, LLC) (32) was used for molecular visualization. Structural models of the variants were generated using MODELLER (version 10.6) (33, 34), with default parameters. The native structure of C5 was modeled based on either the crystal structure with PDB identifier 3CU7 (12) or the AlphaFold-predicted structure (AF-P01031-F1) (35, 36). All structural graphics presented in this article were created using PyMOL based on these models (32).

The index patient experienced two severe episodes of Neisseria spp. meningitis requiring hospitalization, the first in 2018 at the age of 34, and the second in 2023 at age 39 ( Table 1 ). There was no documented history of similar infections during childhood. The initial episode resulted in septic shock secondary to meningococcal infection and bronchopneumonia, necessitating admission to the intensive care unit (ICU) for 20 days. The second episode also presented with clinical features of meningitis and required ICU hospitalization for 14 days. In both instances, the microbiological diagnosis was performed as described in the Materials and Methods section.

Treatment included intravenous antibiotics and dexamethasone for inflammation control when indicated. Following full clinical recovery, the patient was vaccinated against Neisseria meningitidis. Laboratory investigations revealed a marked absence of C5 protein and a severely reduced CH50 activity, consistent with a complement deficiency.

Genetic testing also identified a homozygous variant in the coagulation factor XII gene, associated with increased thrombotic risk. However, the patient was also found to carry the Leu35 variant in factor XIII, which is considered protective against thrombosis, suggesting a low overall thrombotic risk. Additionally, the patient underwent a right hemithyroidectomy and currently reports chronic daily headaches, which are under ongoing clinical evaluation.

Interestingly, the patient’s mother had a documented history of a life-threatening Neisseria infection with sepsis during her youth. One of the patient’s siblings also reported symptoms suggestive of a complement deficiency, including an episode of meningitis at the age of 20 ( Table 1 ). However, no further diagnostic evaluations have been performed to confirm this suspicion. No other relevant symptoms or clinical manifestations related to complement deficiency were identified in the family’s medical history.

The index case and her siblings had received vaccination against Neisseria meningitidis during adolescence, following the Spanish national immunization program. The mother had not been vaccinated previously, likely due to age-related exclusion from the program. Following the diagnosis of meningitis in the index case, the patient, her mother, and all siblings received booster immunization, including vaccination against Neisseria meningitidis and the 20-valent pneumococcal conjugate vaccine, as a precautionary measure.

To investigate the underlying cause of the patient’s recurrent Neisseria spp. infections, molecular analysis of genomic DNA was performed. Two distinct single nucleotide variants (SNVs) in the C5 gene were identified. The first variant, rs567288479 – c.713T>C (p.Ile238Thr), results in the substitution of isoleucine by threonine at residue 238, located in exon 7 within the MG-3 domain ( Figure 1 ). This substitution replaces a nonpolar, hydrophobic amino acid with a polar, uncharged residue. The second variant, c.1949G>T (p.Gly650Val), leads to a change from glycine to valine at residue 650, located in exon 15 within the linker region ( Figure 1 ). This substitution involves replacing a small, flexible amino acid with a bulkier, hydrophobic one, potentially altering local structural conformation.

The p.Ile238Thr variant has been previously reported by Rosain et al. (2017) and El Sissy et al. (2019) (20, 37), and is listed as a variant of uncertain significance (VUS) in public databases (38, 39). In contrast, the p.Gly650Val variant is novel and, to our knowledge, has not been previously described in the literature or variant databases.

Initially, the phase of the variants—whether in cis (on the same allele) or in trans (on opposite alleles) was unknown. To resolve this, targeted Sanger sequencing was performed in close family members, including the patient’s three siblings and mother ( Table 2 ). The results are illustrated in Figure 2 , and the Sanger chromatograms are provided in the Supplementary Material ( Supplementary Figure 1 ).

Interestingly, the patient’s mother carried both variants, raising the possibility of a cis configuration. However, further analysis showed that two siblings (Siblings 1 and 2) carried only one heterozygous variant each, likely inherited from the mother, while a third sibling (Sibling 3) was homozygous for the p.Ile238Thr variant. This segregation pattern strongly supports a trans configuration in both the mother and the patient, indicating that the p.Gly650Val variant was inherited maternally and the p.Ile238Thr variant paternally.

Although the father’s DNA was not available for direct testing due to his passing, the observed inheritance pattern within the family confirms compound heterozygosity in the patient. These findings are summarized in Table 2 and graphically represented in Figure 2 .

To evaluate the functional impact of the identified variants, we performed a functional assays measuring serum levels of complement components, including total hemolytic activity (CH50) ( Table 3 ). C5 and CH50 levels were assessed on at least two separate occasions, and the results were consistent across measurements. The index patient, her mother, and Sibling 3—who is homozygous for the p.Ile238Thr variant—exhibited markedly reduced serum C5 concentrations and significantly decreased CH50 activity. In contrast, Siblings 1 and 2, who are heterozygous carriers of a single variant, showed normal C5 levels and CH50 activity, indicating preserved complement function.

All other measured complement components measured across family members were within normal reference ranges. Notably, C5 and CH50 levels in individuals with clinical manifestations were extremely low or below the limit of quantification (LoQ) of the assay used, supporting a functional C5 deficiency in these individuals.

According to multiple bioinformatic prediction tools, both identified variants were predominantly classified as likely pathogenic, based on their predicted impact on protein structure and function due to the amino acid substitutions ( Supplementary Table 1, 2 ). However, under the criteria established by the ACMG (29), both variants are currently classified as “variants of uncertain significance” (VUS), as they fulfill only one moderate-level criterion (PM2, absence or rarity in population databases).

We next assessed allele frequencies using data from GnomAD (19, 40, 41) and ALFA (20, 41). For the p.Ile238Thr variant, the reported allele frequency in the general population was 0.005112% and 0.04905% in the European (non-Finnish) population according to GnomAD, with no homozygotes individuals reported ( Table 4 ). In the ALFA database, the frequencies were even lower: 0.003% in the general population and 0.004% in Europeans. Based on these data, the p.Ile238Thr variant is considered rare (allele frequency < 0.01%). No allele frequency data was available for the p.Gly650Val variant, as it is not present in either database, supporting its classification as a novel variant. To our knowledge, this is the first documented case of a homozygous p.Ile238Thr variant.

To further assess the structural and functional implications of these mutations, we performed in silico structural modeling using MODELLER, and visualized the resulting models with PyMOL.

For the p.Ile238Thr substitution, the replacement of a nonpolar isoleucine with a polar threonine introduces a hydroxyl group into a hydrophobic region. This disrupts the local environment, as confirmed by comparing the native structure (from PDB 3CU7 and AlphaFold prediction) to our modeled mutant structure. The substitution induces a noticeable structural rearrangement, including the opening of a loop between two antiparallel β-sheets where the residue is located ( Figure 3 ). These changes are predicted to impair protein stability, folding, or secretion, consistent with prior findings in structurally similar C5 variants (5, 42).

Similarly, the p.Gly650Val substitution replaces a small, flexible glycine with a bulkier valine residue. Modeling indicated that the valine side chain creates steric clashes that cannot be accommodated without conformational changes to the protein. Comparison of the native and mutant models revealed a clear structural shift, with displacement of the local loop, upward movement of the residue’s position, and adjustment of adjacent secondary structure elements—including a preceding α-helix and a subsequent β-sheet—to minimize steric interference ( Figure 4 ). As with the p.Ile238Thr variant, these conformational changes likely compromise the protein stability and function, or interfere with secretion.