A single‐center cohort study of patients with hereditary spherocytosis in Central Europe reveals a high frequency of novel disease‐causing genotypes

Hereditary spherocytosis (HS) is the most common red blood cell (RBC) cytoskeleton disorder and the most frequent cause of chronic hemolytic anemia (CHA) in Northern European ancestries.1 HS is phenotypically and genetically heterogeneous, caused by mutations that alter the expression and/or structure of transmembrane, adapter, or spectrin-based cytoskeletal proteins essential for RBC stability, shape, and function.1 These defects lead to the formation of spheroidal, fragile RBCs that are selectively trapped in the spleen, resulting in CHA, jaundice, gallstone formation, and splenomegaly.2 Clinical HS phenotypes range from subclinical to transfusion-dependent severe CHA. Total or near-total splenectomies are curative for severe HS.2 HS-causing defects comprise ANK1 (encoding ankyrin-1), SPTB (encoding β-spectrin), SPTA1 (encoding α-spectrin), SLC4A1 (encoding band 3), and EPB42 (encoding protein 4.2) mutations. HS occurs in all ethnic groups, with differences in overall prevalence and genetic subtypes.2, 3 Next-generation sequencing (NGS) has helped elucidate pathogenic variants underlying HS, the complexity of inheritance modes, and genotype–phenotype correlations in different ancestries including Western Europe,4 North America,5 South America,6 South Asia,7 and East Asia.3, 8 Moreover, better insights into the structures and interactions of key RBC protein complexes were recently provided.9 Here, we report the results of the first Central European single-center study including 113 individuals from 35 families (HS patients, n = 69 including 35 index patients and 34 affected family members; unaffected first-degree relatives, n = 44). Ancestry was Central European in 58/69 patients (86%) and Eastern Turkey (n = 5), Palestine region (n = 4), Iran (n = 1), and Chechnya (n = 1) (Supporting Information S1: Tables S1 and S2). Analyses on patients and both parents were performed in 30/35 (86%) families. Details of diagnostics, clinics, genetic analyses, case reports, and summaries on identified genetic variants are provided in the Supporting Information. In the index group, we identified 34 causative variants in 34/35 patients (97%). Deleterious variants in ANK1 were found in 16/35 (46%), in SLC4A1 in 5/35 (14%), in SPTB in 11/35 (31%) patients, and in SPTA1 in 2/35 (6%) (Table 1 and Supporting Information S1: Figure S25). In one patient (3%), we were unable to identify the genetic cause of HS unequivocally (Supporting Information S1: Patient P35). Details on the localization of the identified variants in the respective proteins are found in Figure 1. Combined annotation-dependent depletion (CADD) versus minor allele frequency (MAF) plots and gDNA variant localization, are found in Supporting Information S1: Figures S6/S13/S20/S23. Of the 34 identified variants, 19 (56%) were not previously reported. Six variants were described previously by us (Table 1), and two of these were subsequently reported in other populations (Supporting Information S1: Families F2/F14). The remaining nine variants were first published elsewhere (Table 1 and Supporting Information S1: Tables S4–S6). From the 34 index patients with pathogenic (n = 30) or likely pathogenic (n = 4) variants, 22 had AD inherited heterozygous variants (64.5%), five had AD de novo heterozygous variants (14.5%), three had autosomal recessive inheritance (one homozygous SLC4A1, two compound heterozygous SPTA1, 9%), and four had variants with unknown inheritance patterns (12%) (Table 1). Most identified variants were high impact (n = 27/34; 79%) including frameshift (n = 14; 41%), nonsense (n = 9, 26%), copy number variants (CNVs, n = 2; 6%), and splice-site variants (n = 2; 6%). The remaining were missense (n = 5; 15%) and intronic variants (n = 2; 6%) (Table 1 and Supporting Information S1: Figure S25). We could demonstrate that the intronic variant in ANK1 (c.1405−9 G>A) generated a new splice–acceptor site resulting in a frameshift and premature stop (Supporting Information S1: Figure S2 and Families F4/F5), as recently reported by others.10 Truncating variants were common in ANK1, SPTB, and SPTA1 (Supporting Information S1: Figures S5/S17/S22). In SLC4A1, 60% of variants were missense (Supporting Information S1: Figures S11 and S12); and two patients with SPTB-associated HS carried missense variants (Supporting Information S1: Figures S17 and S18). One heterozygous AD-inherited CNV in ANK1 (interstitial deletion of exons 27–33) was discovered in Family 8 (Figure 1A and Supporting Information S1: Family F8). The other de novo CNV (132.85 kb large deletion) in Patient P16 affects a total of six contiguous genes (LOC102723729, GPAT4 [OMIM #608143], NKX6-3 [OMIM #610772], ANK1 exons 39–43 [C-terminal region], MIR486-1, and MIR486-2) (Figure 1A and Supporting Information S1: Family F16 and Figures S3 and S4). The 34 affected family members carried the same causative variants as their pediatric index-group relatives (Supporting Information S1: Families F1–F34). In 4/34 index patients (12%) with known cause of disease, ≥2 putative pathogenic variants in HS candidate genes were identified (Supporting Information S1: Families F18/F24/F28/F33 and Figures S9/S15/S16/S21). The low-expression allele Lyon (αLELY), composed of two in cis SPTA1 variants (c.5572C>G and c.6531−12C>T) causing strongly reduced amounts of α-spectrin, was analyzed in the index group, with 18/35 (51%) patients heterozygous for this allele (Table 1 and Supporting Information S1: Families F1–F35).11 The deep intronic “Low Expression allele PRAgue” (α-LEPRA), which enhances an alternative acceptor splice site leading to nonsense-mediated decay,12 was analyzed in all index group patients. Two were identified to carry heterozygous α-LEPRA variants in trans to other pathogenic SPTA1 variants (Table 1 and Supporting Information S1: Families F33 and F34). Overall, causative variants were identified in 68/69 patients (98.5%) and 32/69 (46%) had causative variants in ANK1, 19/69 in SPTB (28%), 14/69 in SLC4A1 (20%), and 3/69 (4%) in SPTA1. Clinical characteristics for all 69 patients are summarized in Supporting Information S1: Tables S1a–e (index group) and S2a–d (affected family members). Data for grading disease severity before splenectomy were available for all 35 index patients and 21 affected family members. Five of the 56 (9%) patients had severe phenotypes (ANK1, n = 2; SPTB, n = 2; homozygous SLC4A1, n = 1). Moderate phenotypes were observed in 34/56 (61%) patients (ANK1, n = 17; SPTB, n = 10; SLC4A1, n = 3; compound heterozygous SPTA1, n = 3; causative variants not unequivocally identified, n = 1) and mild phenotypes in 17/56 (30%) patients (SLC4A1, n = 10; SPTB, n = 5; ANK1, n = 2). There were milder phenotypes in patients with heterozygous SCL4A1 variants, and eosin 5-maleimide binding was highest compared to the other variant cohorts (Supporting Information S1: Table S3 and Figure S1). Hemoglobin levels were significantly higher in patients with heterozygous SLC4A1 variants compared to patients with ANK1 variants (p-adjusted <0.001) and compound heterozygous SPTA1 variants (p-adjusted <0.05) (Supporting Information S1: Table S3 and Figure S1). Of note, hemoglobin levels were higher in patients with SPTB missense variants compared to SPTB stop gain variants (p-adjusted <0.05) (Supporting Information S1: Figure S19). However, the number of patients was too low to allow the establishment of robust phenotype correlation with different types of variants in HS candidate genes in our cohort (Supporting Information S1: Figures S7/S8 [ANK1], S14 [SLC4A1], and S19 [SPTB]). Compared to other studies, we performed all analyses in our hospital and research institute including all available first-degree family members. This allowed us to identify causative variants in 98.5% of patients; and this accuracy is higher than previously reported (e.g., in studies from Western Europe [89%], North America [97%],5 and East Asia [85%]8). We found variants in ANK1 as the most common cause of HS (46%), in line with results from North America (49%)5 and South Asia (53.2%).7 Lower frequencies of causative ANK1 variants were reported in cohorts from Western Europe (27%)4 and East Asia (29%).8 Stop-gain and frameshift mutations were found in 75% of our patients with ANK1 variants, similar to results from East3 and South Asia.7 Interestingly, the intronic variant leading to a frameshift (ANK1, c.1405−9G>A) appears common and was described in different ancestries (Table 1 and Supporting Information S1: Table S4). Variants in SPTB were identified in 31% of index patients, similar in frequency to North America (33%),5 but higher than in Western Europeans (20%).4 The highest incidence of causative SPTB variants was found in Asian HS cohorts.3, 7 Of note, all patients carrying causative variants in SPTB in our study had Central European ancestry. Similar to results from Asia,7 truncating variants were found in most of our SPTB patients. In most populations, <20% of causative defects in SPTB are missense variants.5, 7 In 14% of our index patients, causative variants in SLC4A1 were identified, as observed in West European (14%)4 and North American (13%)5 cohorts. Missense variants were the most prevalent genomic aberrations (60%), corroborating previous findings.3, 4 The patient with the most severe phenotype in our cohort carries a homozygous variant in SLC4A1 (Band3 nullVienna) and we provide an update on the clinical course since the first publication13 (Supporting Information S1: Family F19 and Figure S10). A Western European study identified variants in SPTA1 as the most common cause (36%) of HS.4 Like in other series (South Asia [6.4%],7 East Asia [9%],3 North America [5%]5), 6% of our index patients carried causative SPTA1 variants. Interestingly, all carried the low expression α-LEPRA allele in trans to a high impact SPTA1 variant, and have moderate HS phenotypes. Genotype–phenotype correlations have corroborated that patients who carry heterozygous variants in ANK1 and SPTB (more often high-impact) are phenotypically similar, and patients with heterozygous SLC4A1 variants (more often missense) have milder hematological phenotypes,4, 5, 14 with considerable phenotypical variability in the different genetic subcohorts. Collectively, we provide the first report on the mutational spectrum of HS in a single center in Central Europe. Family analyses, in-house standardized laboratory, and clinical investigations, coupled with NGS analyses, allowed the identification of the genetic cause with high accuracy and confidence in 98.5% of patients and the discovery of unusual causative variants, like deep intronic, splicing, CNVs, and a homozygous pathogenic variant in SLC4A1. We gratefully acknowledge the support of all study participants as well as nurses and treating physicians at St. Anna Children's Hospital Outpatient Clinics. Leo Kager and Kaan Boztug designed the research study. Raúl Jimenez-Heredia, Anna Segarra-Roca, Alexandra Frohne, Karin Nebral, and Petra Zeitlhofer performed the research. Matthias Haimel and Raúl Jimenez-Heredia analyzed the NGS data. Wolfgang Novak and Sebastian K. Eder performed the statistical analysis. Leo Kager, Raúl Jimenez-Heredia, and Kaan Boztug wrote the paper. René Geyeregger, Katharina Roetzer-Londgin, and Oskar A. Haas critically reviewed the paper. Leo Kager reports acting on the advisory board of Agios, Amgen, Bayer, and Novartis (unrelated to this study). The remaining authors declare no conflict of interest. This study was supported through intramural funding of the Ludwig Boltzmann Institute for Rare and Undiagnosed Diseases (LBI-RUD) Vienna, and by a charitable donation of the Kapsch group (www.kapsch.net/kapschgroup) to O. A. H. The data that supports the findings of this study are available in the Supporting Information of this article. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.