Paroxysmal nocturnal hemoglobinuria (PNH) is a rare clonal hematopoietic stem cell disease characterized by a wide spectrum of clinical manifestations, including intravascular hemolysis, a hypercoagulable state, and bone marrow failure. Its global incidence is estimated in 1–1.5 cases per million individuals per year (1).

PNH pathogenesis involves the non-malignant clonal expansion of hematopoietic stem cells with acquired abnormalities in the phosphatidylinositol glycan class A (PIG-A) gene. PIG-A is an X-linked gene that encodes a protein required for glycosylphosphatidylinositol (GPI) anchor synthesis (2). The PIG-A gene comprises six exons and five introns spanning 17 kb (3). Exon 1 (23 bp) is a non-coding 5′-untranslated region, while exon 2 (777 bp) contains half of the coding region. Exons 3 (133 bp), 4 (133 bp), and 5 (207 bp) contain portions of the coding region, and exon 6 (2,316 bp) encompasses the remaining coding region and the 3′-untranslated region. The 5′-flanking region (583 bp) has promoter activity. PIG-A is located on the X chromosome’s short arm (Xp22.1) (2). A single inactivating mutation can produce a PNH phenotype since men have only one X chromosome, and women exhibit X chromosome inactivation (lyonization) in somatic cells, including hematopoietic stem cells (4).

Somatic PIG-A pathogenic variants in hematopoietic stem cells result in partial or complete absence of GPI-linked proteins normally present on the cell membrane. CD55 (decay-accelerating factor) and CD59 (membrane inhibitor of reactive lysis) are essential complement regulatory proteins anchored via GPI. In PNH patients, CD55 and CD59 deficiencies in the clonal blood cells lead to increased red blood cell susceptibility to complement system-mediated cell activaction, cell damage and hemolysis (5, 6).

Clinical observations highlight the heterogeneous nature of PNH patients, with variations in symptom manifestations, hemolysis intensity, and responses to eculizumab treatment. Some patients persistently experience hemolysis, including an extravascular component. A wide array of acquired PIG-A variants has been documented (7), predominantly featuring frameshift mutations (8, 9). Interestingly, these variants do not correlate with clinical factors such as anemia severity or thrombosis presence in PNH patients. This study aims to characterize the mutational profile of 31 Brazilian patients with PNH clones.

Patients harboring PNH clones, identified by flow cytometry, were selected for this study. Flow cytometry analysis was performed using monoclonal antibodies targeting GPI-linked surface antigens like CD59, CD157, and FLAER, as recommended by Borowitz et al. (10). Detection of at least 0.01% of cells lacking GPI-linked marker expression was considered a positive result. The local research ethics committee approved the study, and all patient samples and clinical information were collected following written informed consent.

The patients were classified in three groups according to International PNH Interest Group (11): group 1, classic PNH, characterized by large PNH clones and episodes of intravascular hemolysis; group 2, PNH associated with another hematologic disease, characterized by clinical/laboratory presence of hemolysis in patients who also have another concomitant diagnosis, such as aplastic anemia; and group 3, subclinical PNH, characterized by patients who have small PNH clones without clinical or laboratory manifestations of hemolysis.

A total of 31 patients were included in this study. The median age was 36 years (range 19–72 years), with 55% being male. Twenty patients had classical PNH with large PNH clones (Group 1), three had classical PNH with another hematologic disease (Group 2), and eight had aplastic anemia with small PNH clones (Group 3). The clinical characteristics and laboratory features of the patients are summarized in Tables 2, 3. Among these, 16 (52%) from Groups 1 and 2 experienced hemoglobinuria, and 11 (35%) reported abdominal pain. Twenty-five patients (81%) required transfusion. Eight patients (26%) experienced thrombotic events; seven venous (22%) and one arterial (3%), Including five intra-abdominal venous thromboses (16%), with two cases (6%) being Budd-Chiari syndrome. The median hemoglobin level at sample collection was 9.1 g/dL (range 4.7–14.4 g/dL). The median lactate dehydrogenase (LDH) level was 1,455 U/L (range 159–3,989 U/L, upper limit 480 U/L). Median PNH clone sizes were 79% in monocytes (range 0.03–99.9%), 70% in granulocytes (range 0.01–99.9%), and 37% in red blood cells (range 0.01–95%).

We detected 29 distinct variants in 27 of the 31 patients (refer to Table 4). One mutation had been previously identified: c.55C > T (p.Arg19Trp), a benign polymorphism in exon 2 already reported (Nafa et al., Endo et al.). A c.*7G > T mutation found in the 1,000 Genomes project and two SNPs, c.—427G > C and c.-129A > G, were observed in the general population and are cataloged in the 1,000 Genomes databases. Twenty-five newly identified variants had no prior descriptions in the literature.

Predominantly, point mutations were observed, including 16 variants involving a simple exchange of base pairs, one insertion, 10 small deletion mutations (ranging from one to five base pairs), and one large deletion (19 base pairs). The alterations were categorized as follows: 12 frameshift (41%), 10 missense (35%), two splice site (7%), and one nonsense variant (3%). Four mutations (14%) were found in non-coding regions.

Variants in the PIG-A gene were distributed throughout the gene: three in the 5′ region, 10 in exon 2, two in exon 3, five in exon 4, five in exon 5, one in intron 2, two in intron 4, and one in the 3′ region (Figure 1).

Four patients showed no detectable variants: two with hemolytic PNH (group 1), one with aplastic anemia and PNH clone (group 2) and one with subclinical PNH (group 3).

To the best of our knowledge, this study represents the first comprehensive evaluation of PNH cases using PIG-A gene sequencing in Brazil. We identified 29 somatic variants in 27 out of 31 patients. Among these, only one variant was previously characterized in the literature. Three patients had the c.55C > T variant, initially described by Iida et al. (2). The majority of variants identified (17 out of 29) were deemed likely pathogenic for paroxysmal nocturnal hemoglobinuria (PNH). Six variants have undetermined significance (VUS) and six variants are probably benign.

Since Miyata et al.’s landmark discovery in 1993 (21), which linked PNH pathogenesis to variants in the X-linked PIG-A gene, over a hundred variants have been documented worldwide (9, 22–28). In a smaller cohort, De Carvalho et al. (29) described three Brazilian patients with classical PNH. They used conformation-sensitive gel electrophoresis followed by direct sequencing to identify three distinct variants.

The spectrum of PIG-A variants is broad, predominantly composed of single-nucleotide substitutions, small insertions, and deletions, with large gene deletions being relatively rare. Frameshift mutations often result in premature stop codons, potentially leading to truncated proteins with compromised or lost function. PIG-A variants include nonsense, splice-site variants affecting mRNA processing, and missense that result in single amino acid changes. Some missense variants may confer partial protein functionality, explaining the partial deficiency of GPI-linked proteins in the PNH type II phenotype, while others may lead to a complete loss of function. Our findings resonate with existing literature, wherein frameshift and missense variants are reported frequently; in our study, these accounted for 41 and 35%, respectively. Notably, missense variants with widespread effects are usually situated in pivotal regions; here, four out of nine missense mutations were located in exon 2, which is consistent with previous suggestions that this exon is a critical site for function-affecting variants.

Variants in the PIG-A gene are distributed across the coding region, occurring at random with no significant hotspots, except for a specific sequence in exon 2 reported by some studies (22). The majority of variants were identified in exon 2, likely due to its size and coding significance within the PIG-A gene.

In our study, 19 out of 23 patients with hemolytic PNH exhibited more than one PIG-A variant, suggesting a possible oligoclonal nature of the disease, as previously proposed (8, 30). This theory is supported by the finding that multiple hematopoietic stem cell lineages can acquire distinct PIG-A variants and propagate independently. Genotypic mosaicism within T-cell clones, as demonstrated by Endo et al. (30), underlines the phenotypic diversity observed in PNH. This genotypic diversity within PNH may be due to a selective autoimmune pressure, a hypothesis not yet conclusively proven (31, 32).

Interestingly, patients with concomitant variants did not exhibit clinical differences compared to those with a single mutation. This observation reinforces the idea that PNH is a complex disorder with a multifactorial pathogenesis. In cases of subclinical PNH, no pathogenic variants were identified; only polymorphisms with a high incidence in the general population were present. This finding, coupled with the absence of a sorting method for GPI-deficient leukocytes, suggests that these polymorphisms are likely benign.

The absence of detectable PIG-A variants in four patients, including one with subclinical PNH and three with hemolytic PNH, suggests that mutations may be located in regions not examined by our study, such as non-coding intronic regions. It’s noteworthy that other genes involved in GPI-anchor biosynthesis, like PIG-T and PIG-M, have been implicated in GPI deficiency, but pathogenic variants in these genes are rare and typically associated with autosomal chromosomes, requiring biallelic mutations to manifest (33, 34). These cases often present with a broader clinical spectrum, including developmental delay and epilepsy, which differs from the hemolysis predominant in PNH (35).

The relationship between aplastic anemia and PNH clones, as well as the role of immune selection, provides an interesting angle on the survival advantage of PNH clones (36, 37). This is highlighted by the varying mutation profiles in patients with aplastic anemia and myelodysplastic syndrome compared to those with classical PNH, as reported by Okamoto et al. (38). The diversity of PIG-A variants and their association with different disease phenotypes underscore the complexity of PNH pathogenesis.

Furthermore, the detection of somatic PIG-A variants in individuals with normal hematopoiesis and their prevalence in healthy individuals indicate that these mutations alone are not sufficient for clonal expansion (39–41). The existence of additional genetic abnormalities in PNH patients, such as variants in genes like TET2 and JAK2, suggests a multistep clonal evolution that may contribute to the expansion of PNH clones (42, 43). Additionally, the expression of HMGA2, an architectural transcription factor, has been found to be abnormally high in some PNH patients, potentially contributing to clonal expansion (44).

In terms of clinical manifestations, our study did not reveal a clear correlation between the mutational profile and the severity of hemolytic PNH. This may be due to the wide range of mutations observed. The study also reflects on additional genetic factors, such as BMPR2 and THBD, that may contribute to the risk of thrombosis in PNH patients (45).

In a comprehensive analysis by Chen et al. (46), the clinical implications of variants were elucidated through whole-exome sequencing of genes commonly mutated in PNH, including PIG-A, BCORL1, RUNX1T1, MAP3K4, CSMD1, NOTCH1, FANCD2, PEG3, DIS3, and SETBP1. They found correlations between specific variants and clinical features: the RUNX1T1 mutation was linked to larger PNH clones, higher levels of unconjugated bilirubin, and lower hemoglobin levels; the BCORL1 mutation tended to occur in younger patients; the SRRD mutation was tied to visceral thrombosis; and the EGR4 mutation was associated with myocardial infarction. This emphasizes the potential for targeted mutation analysis to predict clinical outcomes and tailor prophylaxis in PNH patients.

To summarize, the genetic landscape of PNH in Brazilian patients is diverse, with a variety of variant sites and types, and no significant mutational hotspots have been identified. The variants observed were predominantly small deletions and single-nucleotide alterations. Pathogenicity prediction tools suggested that a significant proportion of these variants might play a pathogenic role in the disease: 17 variants were likely pathogenic and five of the six VUS described were probably damaging according to the tools of pathogenicity prediction. In the cohort with hemolytic PNH, a majority (19 out of 23) had variants deemed pathogenic. Interestingly, only polymorphisms, which are common in the general population and likely benign, were detected in patients with subclinical PNH.

The study did not establish a direct correlation between the clinical presentation of hemolytic PNH and specific variants, which may be due to the heterogeneity of the genetic alterations observed. This finding suggests that PNH severity and clinical manifestations are influenced by a complex interplay of factors beyond the PIG-A gene variants alone.

To enhance our understanding of PNH’s heterogeneity, further research involving larger cohorts, including patients from different geographical regions, and a broader analysis of genetic variants, is essential. Such investigations may uncover additional variants implicated in PNH, offering insights into the pathogenesis of the disease and potentially informing the development of personalized therapeutic approaches.