Introduction
Sengers syndrome is a rare autosomal recessive condition. It was first described by Sengers and colleagues in 1975 with the clinical features of congenital cataract, hypertrophic cardiomyopathy, mitochondrial myopathy and lactic acidosis after exercise [
1].
Two forms of this syndrome have been described, a severe neonatal form that causes infantile death and a more benign form with a longer survival into the fourth decade [
2],[
3]. The longer surviving patients had normal developmental milestones. The cause of mortality in Sengers syndrome is heart failure as the result of hypertrophic cardiomyopathy. Van Ekeren et al. investigated 16 clinically-diagnosed patients with Sengers syndrome in order to describe the course of these two forms [
2],[
3]. Although these authors did not observe any histopathologic distinction in muscle or heart samples between the two forms, Perry described 3 cases of neonatal Sengers with significant central nervous system involvement [
4]. The cranial ultrasonography findings in the three siblings included cerebellar hypoplasia and increased echogenicity of the basal ganglia suggestive of calcification. Magnetic resonance imaging demonstrated hypoplasia of both the brainstem and inferior cerebellar vermis, impaired myelination of the cerebral hemispheres and brainstem, and cortical infarction [
5]. These authors suggested that the neonatal form of Sengers syndrome should be described as a mitochondrial encephalomyopathy, because of involvement of the central nervous system. However, the genetic cause of the disease in this family has not been reported. Siriwardena et al. investigated two siblings with Sengers Syndrome and
AGK mutations using magnetic resonance imaging and showed cortical infarction, however in a vascular pattern unlike metabolic strokes that do not respect any arterial distribution [
6].
Recently, Mayr et al. investigated Caucasian Sengers patients from Germany, Italy, The Netherlands and Switzerland and, using whole exome sequencing, identified
AGK as the disease-causing gene in Sengers syndrome [
4].
AGK encodes a mitochondrial transmembrane enzyme, acylglycerol kinase, a multisubstrate lipid with a likely role in cardiolipin biosynthesis. Acylglycerol kinase catalyzes the formation of phosphatidic acid and lysophosphatidic acid [
7] that can participate in phospholipid synthesis or act as signaling molecules regulating a number of cell processes [
7]-[
9]. Cardiolipin plays an important role in structural maintenance of mitochondria and regulating the permeability of the inner membrane. Abnormal mitochondrial morphology has been seen in conditions with cardiolipin impairment, like Sengers and Barth syndromes [
10]-[
12]. Secondary to
AGK mutations, a deficiency of the adenine nucleotide translocator and impairment of ATP synthesis has been reported and seems to play a central role in the pathomechanism of Sengers syndrome [
4],[
13].
We present the clinical features and molecular basis of Sengers syndrome in seven new families of different ethnic origin, documenting for the first time a molecular study of Sengers syndrome patients of non-Caucasian descent.
Discussion
Sengers syndrome is caused by mutations in the
AGK gene [
4], which is located on chromosome 7q34 and consists of 16 exons. To date, only three studies have investigated families with Sengers syndrome and identified different types of loss-of-function mutations in
AGK gene, including start codon mutations (compound heterozygous), nonsense (compound heterozygous), frameshift (compound heterozygous) and splice site mutations (homozygous and compound heterozygous) [
4],[
16]. An overview about the clinical, biochemical and genetic findings of all published and recently identified patients is summarized in Table
1. We investigated the clinical features and molecular basis of Sengers syndrome in seven new affected families. We have identified six novel predicted loss-of-function mutations in
AGK; homozygous c.523_524delAT (p.Ile175Tyrfs*2), c.424-1G > A (splice site), c.409C > T (p.Arg137*) and c.877 + 3G > T (splice site), and compound heterozygous c.871C > T (p.Gln291*) and c.1035dup (p.Ile346Tyrfs*39). Figure
3 provides an overview of the known and newly identified mutations.
To date, less than 10 families with neonatal Sengers syndrome have been reported [
4],[
5]. The average survival of the severe form (with onset in the first month of life), in our and previously published cases, was 4.2 months, whereas Mayr et al. reported patients with the milder form of the disease who survived through their 5
th decade of life (cases PC-5, PC-1 and PC-2, Table
1). Very early mortality (<5 days of life) has also been reported (cases PC-12 and PC-15, Table
1). Calvo et al. described truncating
AGK mutations in two patients with bilateral cataract, severe myopathy and combined complex I, III and IV deficiency [
16]. The first patient (PC-11, Table
1) was born to unrelated parents and harbored compound heterozygous mutations; a splice site mutation c.297 + 2 T > C (p.Lys75Glnfs*12) resulting in a shortened transcript, and a nonsense mutation c.1170 T > A (p.Tyr390*). Her other clinical features included failure to thrive, fatigue, recurrent headaches, osteopenia and premature ovarian failure. She died at the age of 18 months. The second patient (PC-12, Table
1), born to a consanguineous family, was homozygous for a splice site mutation, c.1131 + 1G > T (p.Ser350Glufs*19), that causes a shortened transcript. The infant had severe lactic acidosis (10 17 mmol/L, normal 0.7 2.0) (John Christodoulou, personal communication) and metabolic acidosis, and died at 4 days of age due to cardiorespiratory collapse [
16].
Tachydyspnoea was a feature which was observed only in severe from of the disease (Table
1), and not in any of patients with the milder form. This feature seems to be correlated with poorer prognosis and survival (average ~ 2.8 months).
The unusual features in our patients were nystagmus (Case-1 and Case-5), eosinophilia (Case-3) and cervical meningocele (Case-7). In addition, esotropia was found in two patients (Case-4 and Case-5) with the c.409C > T (p.Arg137*) mutation.
We did not identify a hot-spot region within the
AGK gene showing clustering of mutations; these were equally distributed over the whole gene sequence. So far, all
AGK mutations associated with Sengers syndrome are predicted to be loss-of-function variants. However, we noted that homozygous
AGK nonsense mutations have always resulted in infantile (severe) form of the Sengers syndrome while all patients who survived the first decade harbor at least one splice site variant or a start codon mutation. In fact, two out of the three oldest patients with survival of >35 years carry the c.3G > C (p.Met1?) mutation in combination with two different stop codon mutations and the third one was homozygous for the splice mutation c.1131 + 5G > A, frequently found in patients from The Netherlands. This observation may indicate some AGK activity through residual normal splicing or alternative start codon usage[
17].
Very recently, Aldahmesh et al. identified a splice site mutation in
AGK causing isolated congenital cataract (recessive cataract 38) in a multiplex consanguineous family from Saudi Arabia, thereby extending the mild clinical spectrum associated with
AGK mutations to a non-syndromic form. Urine organic acid profile showed moderately elevated lactate, 3-hydroxyisovaleric, 3-methylglutaric and 3-methylglutaconic acids (Fowzan Alkuraya, personal communication) and interestingly, cardiology evaluation of the affected family members revealed normal results (Table
1). The identified mutation, c.424-3C > G, is predicted to lead to aberrant splicing and predicted premature truncation, p.Ala142Thrfs*4 [
18]. It can be speculated therefore that a small proportion of the normally spliced transcript can still be formed.
Although the role of
AGK mutations in cataractogenesis is unclear, some authors have raised the hypothesis that an impairment of lenticular lipid composition may be of pathophysiological significance in the etiology [
18],[
19]. In our and previously published cases with the severe form of Sengers syndrome, cataract was present at birth or detected shortly thereafter, whereas in milder form, it can remain undiagnosed for several months (18 months of age in PC-3, Table
1).
Several patients (Case-7, Case-8, PC-3 and PC-18, Table
1) with a milder form of Sengers syndrome did not develop lactic acidosis, whereas in the severe form, none of the patients had normal levels of lactic acid. Siriwardena et al. reported a Sengers patient with the start codon mutation, p.Met1?, who interestingly, at the age of 3 years, did not have skeletal myopathy. He was able to run and play without any limitation and his creatine phosphokinase levels were also normal. His skeletal muscles biopsy, at the age of 4 months, did not show presence of COX negative fibers [
6]. However, his sibling, obviously with the same
AGK mutation, was affected with the severe form of the diseases and died at the age of 6 months. These observations suggest that myopathy can develop independent from lactic acidosis and that the absence of lactic acidosis favors a better prognosis with longer-term survival, even in patients with the same
AGK mutations. This observation resembles patients with Barth syndrome, which is known to show clinical variability found in siblings with identical
TAZ genotype [
20]. Remarkably, these two mitochondrial disorders also seem to result in a decrease of complex I, which has been demonstrated by western blot analysis in Barth syndrome [
21] and shown here in a heart and to milder extent a skeletal muscle sample (Figure
4).
Based on the finding of severe mtDNA depletion in skeletal muscle in two patients reported by Calvo and colleagues [
16], Sengers syndrome is now also called cardiomyopathic mitochondrial DNA depletion syndrome-10 (OMIM 212350). In contrast to this published study, we did not find any indication for mtDNA depletion in muscle or heart from three patients with samples available for testing. Moreover, in the original publication of the clinical syndrome by Sengers and colleagues [
1], the biochemical analysis of mitochondrial enzyme complex activities revealed normal results, which is incompatible with a severe loss of mtDNA copy number [
1],[
6]. To date, only nine patients with
AGK mutations have been reported to have a combined respiratory chain deficiency indicating that mtDNA depletion is not a common feature of Sengers syndrome.
Based on our findings, we can describe two distinct forms of Sengers syndrome, an infantile or severe form and a mild form. The infantile form of Sengers syndrome is caused by homozygous AGK nonsense mutations and is characterized by early onset (within the first few months of life) cardiomyopathy and lactic acidosis that causes death in infancy. Some patients who carry at least one AGK splice site variant or a start codon mutation develop a milder form of Sengers syndrome and have a markedly better prognosis. These patients usually develop cardiomyopathy in later stages. They can survive even through their 5th decade of life and in the extreme case, they may only develop cataract.
Currently, no treatment is available for this fatal condition; therefore it is important to identify carriers in at-risk families in order to provide genetic counseling and prenatal diagnosis.
In summary, we have identified six novel AGK mutations causing Sengers syndrome. Our findings, along with the other studies, indicate clinical heterogeneity in patients harboring AGK mutations. However, interestingly, congenital or early-onset cataract is the common phenotypic finding in all patients, suggesting a role for AGK in the homeostasis of the human eye lens. We also suggest that infants with cataract, even in the absence of cardiomyopathy, should be screened for Sengers syndrome or AGK mutations. This study verifies the causative role of AGK in Sengers syndrome and expands the genotype-phenotype correlations of mutations in this gene. Our results have important applications in genetic counseling and prenatal diagnosis.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
AH and HP contributed as project leads across the whole project. AH: Study design, literature search, data collection, data analysis, interpretation of results, figures, and writing the first draft of manuscript. TH: Sequencing, data analysis and interpretation of results, and contribution to the manuscript and figures. UA and EH: Sequencing, data analysis and interpretation. RGF and JAM: Protein and mtDNA analysis, interpretation of results, figures, and contribution to the manuscript and figures. MA, HM, HH-K, RAB, RGF, AR, A-SL, TK, AD, NP, MAS, NSG, PFC, RWT: patient recruitment and clinical examination, data collection, and contribution to the manuscript. HP: Study design, data analysis, interpretation of results, figures, and contribution to the manuscript and figures. All authors reviewed the manuscript for important intellectual content and approved the final manuscript.