Abstract
Background and objectives Atypical hemolytic uremic syndrome is characterized by vascular endothelial damage caused by complement dysregulation. Consistently, complement inhibition therapies are highly effective in most patients with atypical hemolytic uremic syndrome. Recently, it was shown that a significant percentage of patients with early-onset atypical hemolytic uremic syndrome carry mutations in diacylglycerol kinase-ε, an intracellular protein with no obvious role in complement. These data support an alternative, complement-independent mechanism leading to thrombotic microangiopathy that has implications for treatment of early-onset atypical hemolytic uremic syndrome. To get additional insights into this new form of atypical hemolytic uremic syndrome, the diacylglycerol kinase-ε gene in a cohort with atypical hemolytic uremic syndrome was analyzed.
Design, setting, participants, & measurements Eighty-three patients with early-onset atypical hemolytic uremic syndrome (<2 years) enrolled in the Spanish atypical hemolytic uremic syndrome registry between 1999 and 2013 were screened for mutations in diacylglycerol kinase-ε. These patients were also fully characterized for mutations in the genes encoding factor H, membrane cofactor protein, factor I, C3, factor B, and thrombomodulin CFHRs copy number variations and rearrangements, and antifactor H antibodies.
Results Four patients carried mutations in diacylglycerol kinase-ε, one p.H536Qfs*16 homozygote and three compound heterozygotes (p.W322*/p.P498R, two patients; p.Q248H/p.G484Gfs*10, one patient). Three patients also carried heterozygous mutations in thrombomodulin or C3. Extensive plasma infusions controlled atypical hemolytic uremic syndrome recurrences and prevented renal failure in the two patients with diacylglycerol kinase-ε and thrombomodulin mutations. A positive response to plasma infusions and complement inhibition treatment was also observed in the patient with concurrent diacylglycerol kinase-ε and C3 mutations.
Conclusions Data suggest that complement dysregulation influences the onset and disease severity in carriers of diacylglycerol kinase-ε mutations and that treatments on the basis of plasma infusions and complement inhibition are potentially useful in patients with combined diacylglycerol kinase-ε and complement mutations. A comprehensive understanding of the genetic component predisposing to atypical hemolytic uremic syndrome is, therefore, critical to guide an effective treatment.
Introduction
Hemolytic uremic syndrome (HUS) is a rare, life-threatening disease characterized by thrombocytopenia, hemolytic anemia, and acute renal failure (1). The most frequent form of HUS follows a diarrheal prodrome and is associated with infections involving shiga toxin–producing Escherichia coli strains. Five to ten percent of patients with HUS lack an association with this type of infection. This atypical form of HUS (aHUS) has the poorest long-term prognosis; it is characterized by recurrences and presents a mortality rate approaching 30% (1,2). Genetic analyses in patients with aHUS have revealed that most cases associate with mutations and polymorphisms in complement genes and that the disease develops as a consequence of defective protection of cellular surfaces from complement activation (3–17). The recognition that aHUS is a disorder involving complement-dependent tissue damage provided strong support for the implementation of aHUS therapies based in complement inhibitors (18).
In contrast to this complement-related aHUS, Lemaire et al. (19) have reported that as many as 27% of aHUS cases presenting in the first 1 year of life are caused by the deficiency of diacylglycerol kinase-ε (DGK-ε) encoded by the diacylglycerol kinase-ε (DGKE) gene. It has been proposed that lack of DGK-ε causes enhanced signaling through AA-containing DAGs and results in a prothrombotic phenotype (19,20). Because these patients lacked discernible complement alterations, it was suggested that DGKE-associated aHUS represents an alternative mechanism leading to thrombotic microangiopathy that is independent of complement dysregulation (19).
To get additional insights into this new form of aHUS and estimate the prevalence of carriers of DGKE mutations in the aHUS Spanish cohort, we performed analysis of the DGKE gene in 83 patients with a disease onset in the first 2 years of life.
Materials and Methods
Patients
Our aHUS cohort includes patients, mainly of Spanish origin, enrolled since 1999. We have performed complement studies that have allowed us to identify mutations, risk haplotypes, and autoantibodies in these patients, most of which have already been published. Currently, our cohort includes 289 patients with aHUS: 262 patients from Spain, 9 patients from other European countries, 4 patients from north Africa (Morocco, Tunisia, and Senegal), 6 patients from the United States, and 8 patients from South America. All these patients were diagnosed at specific centers in the countries mentioned and subsequently submitted to the Spanish aHUS registry for complement functional and genetic analyses. aHUS was diagnosed on the basis of microangiopathic hemolytic anemia and thrombocytopenia defined by an hematocrit <0.3 (30%), hemoglobin level <10 g/dl, serum lactate dehydrogenase level >460 units/L, undetectable haptoglobin level, fragmented erythrocytes in the peripheral blood smear, and platelet count <150,000/μl associated with acute renal failure. The studies reported here have the approval of the Institutional Review Board. Informed consent was provided by all individuals participating in the study according to the Declaration of Helsinki.
Complement Profile Assessment
Serum concentrations of C3 and C4 were evaluated by nephelometry (Siemens, Marburg, Germany). Factor H (FH) and factor I (FI) serum levels were measured by ELISA (21,22). Membrane cofactor protein (MCP; CD46) levels on peripheral blood cells were determined by flow cytometry (8). The presence of anti-FH antibodies and C3 nephritic factor and the functional analysis of FH using the sheep erythrocytes hemolysis assay were performed as previously described (23–25).
Mutation Screening and Genotyping
Exons of the genes encoding factor H (CFH), membrane cofactor (MCP), factor I (CFI), C3 (C3), factor B (CFB), and thrombomodulin (THBD) were amplified from genomic DNA using primers derived from the intronic sequences as described (5,10,11,13,26). Exons of DGKE were amplified with the primers described in Supplemental Table 1. Automatic sequencing was performed in an ABI3730 sequencer using a dye terminator cycle sequencing kit (Applied Biosystems). Copy number variations and genomic rearrangements were assessed by multiplex ligation-dependent probe amplification and custom-designed high-density 8×15,000 oligonucleotide CGH arrays spanning the RCA gene cluster (median resolution=110 bp; AMADID 040193; Agilent Technologies, Santa Clara, CA) (27). CFH and MCP genotyping was performed as described previously (8).
In Silico Analyses of Mutations
The probability of a genetic variant resulting in structural or functional alterations was calculated using bioinformatics prediction tools that discriminate neutral polymorphisms from amino acid substitutions of likely functional importance. To minimize the possibility of false-positive or false-negative results, we have applied four different computational algorithms publicly available: Sorting Intolerant from Tolerant (28) (SIFT; http://sift.jcvi.org), Polymorphism Phenotyping (29) (PolyPhen; http://genetics.bwh.harvard.edu), Mutation Taster (30) (http://www.mutationtaster.org), and Align-GVGD (31) (http://agvgd.iarc.fr). The SIFT algorithm predicts whether an amino acid substitution affects protein function on the basis of sequence homology among related genes and domains over evolutionary time and the physicochemical properties of the amino acid residues. PolyPhen also incorporates the analyses of sequence conservation and the nature of the amino acid residues involved as well as the location of the substitution within the structure of the protein. By accessing a variety of heterogeneous biologic databases and analytical tools, Mutation Taster is able to identify the missense mutations most likely to have functional effects, such as changes to the transcriptional level and pre-mRNA splicing. Align-GVGD combines the biophysical characteristics of amino acid and protein multiple sequence alignments to predict where missense substitutions in genes of interest fall in a spectrum from enriched deleterious to enriched neutral.
Results
The Spanish aHUS cohort includes 83 patients enrolled between 1999 and 2013 with a disease onset in the first 2 years of life; 62 of these patients do not have a discernible genetic or autoimmune abnormality that helps to explain the development of the disease. We performed the analysis of the DGKE gene and found that one of these patients (HUS299) was a homozygote for a frameshift DGKE mutation. We also screened 21 patients with early-onset aHUS with identified genetic factors and surprisingly, found 3 additional patients belonging to two unrelated pedigrees carrying compound heterozygous mutations in DGKE (Figures 1 and 2, Table 1). We did not find heterozygote carriers of DGKE mutations in our aHUS cohort.
Pedigrees of patients with atypical hemolytic uremic syndrome carrying diacylglycerol kinase-ε (DGKE) mutations. For each pedigree, the index patients are indicated with arrows, and the chromatograms for the corresponding mutations are shown. Segregation of the mutations is indicated with colored symbols. HUS, hemolytic uremic syndrome; ND, genetic analyses have not yet been performed.
Location of the diacylglycerol kinase-ε mutations found in Spanish patients with atypical hemolytic uremic syndrome (aHUS). A diagram of the diacylglycerol kinase-ε protein depicting the different protein domains is shown. Arrows identify the positions of the five mutations found in the screening of the early-onset Spanish patients with aHUS. C1 domain, kinase catalytic domain and kinase accessory domain; HD, hydrophobic domain.
Diacylglycerol kinase-ε and complement gene mutations found in patients with early-onset atypical hemolytic uremic syndrome
Patients with Isolated DGKE Mutations
HUS299 is a 4-year-old boy from Morocco who presented with aHUS at 13 months, coincident with a diarrhea episode. He required peritoneal dialysis for 22 days but evolved well without recurrences. Three years after remission, the patient remains asymptomatic, with only a residual microalbuminuria (Table 2, Supplemental Material). HUS299 is homozygous for a frameshift mutation in exon 12 (c.1608_1609del; p.H536Qfs*16) of DGKE that results in a truncated DGK-ε protein lacking the last 31 amino acids (Figures 1 and 2, Table 1), which is very likely deleterious. This novel DGKE genetic variation was not found in 80 control chromosomes from North African populations, and it is not described in the available databases. Additional genetic analysis in this patient failed to reveal other genetic abnormalities. He was also negative for anti-FH autoantibodies, and his plasma complement protein levels, including C3, were normal (Table 3).
Clinical characteristics of the patients with atypical hemolytic uremic syndrome carrying diacylglycerol kinase-ε mutations
Complement profiles in patients with atypical hemolytic uremic syndrome carrying mutations in diacylglycerol kinase-ε
Patients with Concurrent Mutations in DGKE and THBD
HUS40 and HUS39 are siblings from an Argentinean family of European ancestry. HUS40 presented with HUS at 7 months of age. Five months later, she had a recurrence that required hemodialysis and continuous ambulatory peritoneal dialysis for a few months. She evolved satisfactorily, partially recovering renal function without dialysis requirement. A third recurrence occurred at 2 years of age that responded well to a protocol of daily plasma infusions, which were suspended 1 year later. She did not suffer any more recurrences. She is now 17 years old and clinically well, with normal blood parameters, serum creatinine of 1.5 mg/dl, and Ccr of 41 ml/min per 1.73 m2 (Table 2, Supplemental Material). Her brother, HUS39, presented with aHUS at 3 months. He was anuric, requiring acute peritoneal dialysis. On the basis of the experience with his sister (HUS40), he was immediately treated with daily plasma infusions. After 10 days, his renal function improved, and he was discharged, maintaining the daily plasma infusions for 2 months (10 ml/kg per day). Interestingly, when plasma infusions were spaced biweekly, he had an aHUS recurrence that responded well when plasma infusions were returned to daily basis. On the basis of his clinical improvement, plasma infusions were again spaced progressively and finally suspended 5 years later. He is now 11 years old and clinically well, with normal blood parameters, serum creatinine of 0.8 mg/dl, and Ccr of 57 ml/min per 1.73 m2 (Table 2, Supplemental Material).
HUS40 and HUS39 are compound heterozygotes for a nonsense mutation (c.966G>A; p.W322*) in exon 6 and a missense mutation (c.1493C>G; p.P498R) in exon 11 of the DGKE gene (Figures 1 and 2, Table 1). The p.W322* mutation has been described previously as a pathogenic mutation associated with a rare European haplotype (19). Interestingly, the p.W322* mutation in our patients is also associated with the same DGKE haplotype (Supplemental Table 2). The second mutation p.P498R is a novel genetic variation that results in an amino acid substitution in exon 11 of the DGKE gene, a region that encodes the kinase accessory domain (Figure 2). p.P498R was not found in 140 normal control chromosomes and has not been reported previously in the available DGKE mutation databases. We lack assays to test the functionality of the DGKE mutations and have no biopsy material from the patients to test DGK-ε expression. However, a very strong argument supporting that p.P498R is a DGKE deleterious mutation is that the patients carrying this mutation presented it together with the deleterious p.W322* mutation (patients are compound heterozygotes), which is in agreement with the recessive model of inheritance described for pathogenic DGKE mutations. It would be extremely unlikely that this pairing of mutations would have occurred by chance if such infrequent genetic variants were not pathogenic. Furthermore, we have used bioinformatics prediction tools that discriminate neutral polymorphisms from amino acid substitutions of likely functional importance to support that the identified DGKE changes represent true disease mutations (Materials and Methods). These in silico analyses indicate that p.P498R is a damaging mutation (Table 4).
In silico analyses of missense mutations
Genetic analyses in these patients revealed that both patients are also heterozygotes for a mutation in THBD (c.1456G>T; p.D486Y) previously found associated with aHUS and shown to be pathogenic (16) (Figure 1). CFH and MCP genotyping showed that HUS39 and HUS40 are heterozygotes for the MCP and CFH aHUS risk haplotypes, respectively. The search for anti-FH autoantibodies was negative, and the complement plasma levels were normal, although C3 levels in these individuals, particularly HUS40, were in the lower part of the normal range (Table 3).
Patients with Concurrent Mutations in DGKE and C3
HUS272 is a 4-year-old girl from a German–Spanish family who presented with aHUS at 8 months of age, coincident with an upper respiratory tract infection. She had periorbital edema, low plasma total protein levels, hematuria, and nephrotic proteinuria (Table 2). After 48 hours, she developed generalized edema, hypertension, and thrombocytopenia, and schistocytes were observed at the peripheral blood smear. A biopsy supported the diagnosis of aHUS. She evolved favorably with a progressive decrease of proteinuria levels. Eight months later, coincident with a vaccination, she suffered an episode of oliguria and edema associated with increased proteinuria, which resolved by raising the Enalapril doses. However, 3 months later, after an acute upper tract infection, she had an aHUS recurrence, and her clinical situation deteriorated. C3 was slightly below the normal range. She initiated biweekly plasma infusions, which led to stabilization of the clinical symptoms and near normalization of all blood parameters (Figure 3); however, proteinuria and gross hematuria persisted. One year later, C3 levels dropped to 0.77 g/L, the edema returned, and her clinical situation worsened again. It was decided to switch her from the biweekly plasma infusions to Eculizumab treatment, and thereafter, her clinical situation improved significantly. However, a urine sample revealed ongoing proteinuria. She is currently treated biweekly with Eculizumab, and the clinical situation had improved. Previously, she always had peripheral edema when she had infections; under Eculizumab treatment, however, she remained without obvious edema (Table 2, Supplemental Material).
Evolution of blood parameters in hemolytic uremic syndrome 272 (HUS272). The evolution of the (A) C3, (B) platelets, and (C) lactate dehydrogenase (LDH) serum levels in patient HUS272 from the disease onset are shown. Arrows indicate the start of the plasma therapy (fresh frozen plasma [FFP]) and the Eculizumab (Ecu) treatment. C3 levels in the mother (Mo) and father (Fa) are also indicated.
Genetic analysis showed that HUS272 is a compound heterozygote for a nonsense mutation (c.1452del; p.G484Gfs*10) in exon 11 of DGKE that truncates the last 69 amino acids of the protein (most likely deleterious) and a missense mutation (c.744G>C; p.Q248H) in exon 4 of the DGKE gene, encoding the kinase catalytic domain of DGK-ε (Figures 1 and 2, Table 1). p.G484Gfs*10 and p.Q248H are novel mutations that were not found in 140 normal control chromosomes and have not been reported previously in the available DGKE mutation databases. Similar to the case of the p.P498R mutation found in HUS40 and HUS39, the concurrence of p.G484Gfs*10 and p.Q248H in HUS272 and the results of their analysis in silico strongly support that these DGKE mutations are pathogenic (Table 4).
Critically, the analyses of all known aHUS genetic risk factors showed that HUS272 carries, in addition to the DGKE mutations, a novel mutation in the C3 gene (c.784G>T; p.G262W) (Figure 1) that, in silico analysis, is also clearly identified as a damaging mutation (Table 4). p.G262W was not found in 140 normal control chromosomes and is not included in genetic variation databases. Interestingly, this patient carries the C3 mutation combined with an MCP risk haplotype in homozygosity (Table 3), which is characteristic of patients with aHUS carrying gain-of-function mutations in C3 and CFB (13,32). Complement plasma levels were normal, although C3 levels in this patient were always below or in the lower part of the normal range (Table 3). Notably, her mother, who is also a carrier of the p.G262W C3 mutation, also presents decreased plasma C3 levels (Figure 3A).
Discussion
Previous studies have identified two groups of patients with mutations in the DGKE gene. Ozaltin et al. (20) reported nine patients from three consanguineous pedigrees presenting with membranoproliferative GN with histologic signs of both glomerular microangiopathy and endothelial distress at ages ranging from 0.8 to 17 years old. No complement data were reported for these patients (20). A second study by Lemaire et al. (19) reported 13 patients from nine unrelated pedigrees also carrying mutations in DGKE. In contrast, all these patients presented within the first 1 year of life with aHUS. Notably, these patients belong to a group of pediatric-onset aHUS that was selected to include only patients in whom mutation in known aHUS-associated genes or anti-FH autoantibodies has not been identified (19). Here, we have analyzed the DGKE gene in our aHUS cohort and found that 5% (4 of 83) of our early-onset patients carry DGKE mutations, including both carriers and noncarriers of mutations in other aHUS-associated genes.
Disease presentation and evolution in our four patients was comparable with that of the patients carrying DGKE mutations described by Lemaire et al. (19) These patients met the clinical criteria for aHUS at presentation and had disease recurrences within early childhood. Later, one patient clearly progressed to chronic glomerulopathy with hematuria, proteinuria, hypertension, and renal failure, which is distinctive of DGKE mutation carriers from other patients with aHUS (Table 2).
Interestingly, we found that three of our patients also carry mutations in other known aHUS-associated genes. Clinical course and severity of the disease in these patients were different from those in the patient with isolated DGKE mutations. HUS39 and HUS40, who also carry a pathogenic heterozygous mutation in THBD, had various aHUS recurrences shortly after the disease onset that responded well to daily plasma infusions. Notably, the course of the disease was more benign in the affected brother (HUS39), who was intensively treated with plasma immediately after the onset of the disease (Table 2). These data point to a potential beneficial effect of plasma infusions in our patients and argue against previous remarks (on the basis of data from patients carrying exclusively loss-of-function DGKE mutations) not supporting the use of plasma in affected carriers of DGKE mutations (19). Notably, it has been reported that most (88%) patients with aHUS associated with THBD mutations responded well to plasma therapy (33). Therefore, we suggest that the association of a THBD mutation with the DGKE mutations in HUS39 and HUS40 results in a potentially more severe phenotype because of an increased frequency of aHUS recurrences caused by the THBD mutation. Consistent with other patients with aHUS carrying THBD mutations, our patients responded well to plasma therapy.
HUS272, who in addition to the DGKE mutations, also carries a likely pathogenic heterozygous mutation in C3 and the MCP risk haplotype in homozygosity, presented several aHUS recurrences shortly after onset. Biweekly plasma infusions were effective in normalizing blood parameters, and subsequent Eculizumab treatment resolved the infection-associated edemas that were typical in this patient (Figure 3). Compared with the other patients with DGKE mutations in this case series, the association of a C3 gene mutation in this particular patient possibly contributed to more severe disease with chronic activation of thrombotic microangiopathy despite plasma treatment and more importantly, a positive response to complement C5 blockage with Eculizumab, which was conducted to remission.
Finally, HUS299, carrying exclusively DGKE mutations, recovered renal function without plasma infusions or Eculizumab treatment and has remained stable without recurrences since the aHUS onset 3 years ago (Table 2).
These findings support a potential role of mutations in complement genes in patients carrying DGKE loss-of-function mutations. The novelty of our approach is searching for DGKE mutations in patients with aHUS with and without mutations in known aHUS-associated complement genes. It is remarkable that, after screening 83 patients with aHUS with a very early onset (<2 years) without excluding those patients with previously identified mutations in known aHUS-associated genes (21 patients), we found 3 patients carrying recessive DGKE mutations who also carry mutations in other known aHUS-associated genes and just 1 patient with isolated DGKE mutations. These data and the evident correlation of disease recurrences with infection in our patients (Supplemental Material) suggest that the onset and severity of the chronic renal pathology associated with DGKE mutations may be influenced by additional complement-related aHUS risk factors. If these data are replicated in other aHUS cohorts and a role for complement dysregulation is finally established in DGKE-related aHUS, the suggestion that current treatment on the basis of plasma therapy and complement inhibition are not beneficial in individuals with DGKE mutations should, perhaps, be reconsidered, at least for patients with coexistence of DGKE and complement gene mutations. Preventing complement-related thrombotic microangiopathy episodes may result in a more benign course of the renal disease in carriers of DGKE mutations.
In conclusion, we found that 5% of the patients in our aHUS cohort who had a very early onset of the disease (<2 years) carry recessive mutations in DGKE. Although this percentage is small compared with 27%, which was reported for the French cohort, it is still a significant number. Our patients carry a total of five different mutations in the DGKE gene. Four of them are novel mutations, and one (p.W322*) mutation seems to be a relatively prevalent mutation of European origin. The disease presentation and evolution in our patients was comparable with that of patients carrying DGKE mutations described by Lemaire et al. (19) in the French aHUS cohort. However, in contrast with that previous study, we found that three of our patients carry additional mutations in the known aHUS candidate genes C3 and THBD. Although additional analyses in other aHUS cohorts are needed to replicate our findings, these data suggest that complement-mediated aHUS episodes have a role in modulating the onset and severity of renal disease in carriers of DGKE mutations. Finally, this work also illustrates that a comprehensive understanding of the genetic component predisposing to the pathology is critical to guide appropriate diagnostics and effective treatment in aHUS.
Disclosures
B.H., M.A., L.L., and S.R.d.C. have received honoraria from Alexion Pharmaceuticals for giving lectures and participating in advisory boards. None of these activities has had any influence on the results or interpretation in this article. The remaining authors declare no conflicts of interest.
Acknowledgments
We thank the patients and their relatives for their participation in this study. We also thank Dr. Gema Ariceta for comments and suggestions during the preparation of this manuscript. We thank Drs. J. Perez and L. Olavarrieta from Secugen SL for invaluable help in the design for the diacylglycerol kinase-ε sequencing strategy.
Work in this report was funded by Spanish “Ministerio de Economía y Competitividad” Grants SAF2011-26583, PS09/00122 (to M.L.T.), SAF2012-38636 (to M.L.T.), and PI1200597 (to P.S.-C.) and Fundación Renal Iñigo Alvarez de Toledo and Seventh Framework Programme European Union Project EURenOmics Grant 305608 (to S.R.d.C.). In addition, this work was supported by Autonomous Region of Madrid Grant S2010/BMD-2316 (to M.L.T., P.S.-C., and S.R.d.C.).
Footnotes
Published online ahead of print. Publication date available at www.cjasn.org.
This article contains supplemental material online at http://cjasn.asnjournals.org/lookup/suppl/doi:10.2215/CJN.01640214/-/DCSupplemental.
See related editorial, “Factors Influencing Treatment of Atypical Hemolytic Uremic Syndrome,” on pages 1516–1518.
- Received February 12, 2014.
- Accepted May 12, 2014.
- Copyright © 2014 by the American Society of Nephrology