Article Text
PDF
Abstract
Background Myelodysplasia, infection, restriction of growth, adrenal hypoplasia, genital phenotypes and enteropathy (MIRAGE) syndrome is a recently described congenital disorder caused by heterozygous SAMD9 mutations. The phenotypic spectrum of the syndrome remains to be elucidated.
Methods and results We describe two unrelated patients who showed manifestations compatible with MIRAGE syndrome, with the exception of haematological features. Leucocyte genomic DNA samples were analysed with next-generation sequencing and Sanger sequencing, revealing the patients to have two de novoSAMD9 mutations on the same allele (patient 1 p.[Gln695*; Ala722Glu] and patient 2 p.[Gln39*; Asp769Gly]). In patient 1, p.Gln695* was absent in genomic DNA extracted from hair follicles, implying that the non-sense mutation was acquired somatically. In patient 2, with the 46,XX karyotype, skewed X chromosome inactivation pattern was found in leucocyte DNA, suggesting monoclonality of cells in the haematopoietic system. In vitro expression experiments confirmed the growth-restricting capacity of the two missense mutant SAMD9 proteins that is a characteristic of MIRAGE-associated SAMD9 mutations.
Conclusions Acquisition of a somatic nonsense SAMD9 mutation in the cells of the haematopoietic system might revert the cellular growth repression caused by the germline SAMD9 mutations (ie, second-site reversion mutations). Unexpected lack of haematological features in the two patients would be explained by the reversion mutations.
- adrenal disorders
- genetics
- haematology (incl Blood Transfusion)
- mirage syndrome
- reversion mutation
Statistics from Altmetric.com
Introduction
MIRAGE syndrome (OMIM #617053) is a recently described multisystem disease entity caused by heterozygous SAMD9 mutations that mostly occur de novo.1 The term ‘MIRAGE’ is an acronym for the six core features (myelodysplasia, infection, restriction of growth, adrenal hypoplasia, genital phenotypes and enteropathy). SAMD9 locates on chromosome 7q21.2 and encodes a 1589 amino acid protein that acts as a suppressor of cell growth.2 MIRAGE syndrome-associated SAMD9 mutations are supposed to be activating mutations, because expression of mutant proteins causes profound growth repression in vitro, whereas the wild type (WT) protein causes only mild growth restriction.1 3 SAMD9 has also been reported to cause, when inactivated, normophosphatemic familial tumorous calcinosis (NFTC; OMIM #610455).4 SAMD9 mutation-carrying patients with NFTC have only been described among Jewish Yemenite, and the relation between the mutations and the phenotype is less clear.
Except for NFTC, 22 individuals with a heterozygous SAMD9 mutation have been reported to date.1 3 5 Most mutation carriers have five or six core MIRAGE features, thereby forming a recognisable disease entity. An exception is a family with the p.Glu1136Gln mutation,5 in which three mutation-carrying siblings had haematological manifestations including myelodysplastic syndrome. Although the haematological phenotype seemed typical for MIRAGE syndrome, the siblings did not have most of other core MIRAGE features.5 Furthermore, the mother of them had the identical mutation but was unaffected. These reports suggest that the full clinical spectrum of SAMD9 defects remains to be determined.
Here, we describe two unrelated patients with molecularly confirmed MIRAGE syndrome with unusual lack of haematological phenotype. They both had an additional SAMD9 mutation in leucocyte DNA, likely explains the atypical clinical picture.
Materials and methods
Patients
Full description of clinical information is available in the online supplementary file 1. Patient 1 was born preterm with a birth weight of 1058 g (−3.3 SD) to Japanese parents. Skin hyperpigmentation, hypotension and hypoglycaemia were noted soon after the birth. She was diagnosed as having adrenal insufficiency based on hormonal measurements and has been treated with hydrocortisone and fludrocortisone. Adrenal glands were not visible by ultrasonography. She had multiple skeletal malformations including scoliosis and joint contracture in her wrists and ankles (figure 1A). She had complete female-type external genitalia despite the 46,XY karyotype. She had multiple episodes of opportunistic infection from early infancy, with decreased serum levels of IgG and IgM (see online supplementary figure S1). Due to severe gastro-oesophageal reflux and chronic diarrhoea, she was continuously fed via a gastric tube, and gastrostomy was performed at age 5 years. She had severe developmental delay: she could neither say words nor control her head. Cranial CT scan revealed hydrocephalus (figure 1A). Hypolacrima and corneal ulcers were noted. The growth was severely restricted (figure 1B). With the exception of an episode of disseminated intravascular coagulation secondary to stomach perforation and surgical correction at age 1 year, she had no history of blood transfusions. From 8 years of age, she had renal tubular acidosis, glucosuria, defects in phosphate reabsorption and urinary concentration, with high urine β2 microglobulin excretion.
Supplementary file 1
Clinical findings of the two patients. (A) Radiological findings of patient 1: she had multiple skeletal malformations including scoliosis and joint contractures in her ankles. CT scan revealed hydrocephalus. (B) Ethnicity and gender-matched growth charts of the two patients. Patient 1 was born small and has shown continuous growth restriction during her postnatal life. In contrast, patient 2 showed catch-up growth until age 8 years. After administration of supraphysiological steroid dosages, her growth stagnated. On therapy with 0.035 mg/kg/day growth hormone started at age 10.8 years, she showed catch-up growth.
Patient 2 was born preterm with a birth weight of 1870 g (−2.3 SD) to German parents. She was diagnosed as having oesophageal achalasia at 5 months and received surgery including Heller’s myotomy. Due to vomiting, gastro-oesophageal reflux and aspiration, she required gastrostomy and jejunal tube feeding. She presented periodic episodes of fever, diarrhoea, cramping abdominal pain, vomiting and hypertension every 2–3 weeks. At this point, serum cortisol measurements were normal. At 3 years of age, proteinuria was found, and she was diagnosed with interstitial nephritis and mainly selective glomerular proteinuria. The cognitive development was considered to be in the low normal range, although development of speech was distinctly retarded. She had slightly decreased IgG levels, but IgA and IgM were high (see online supplementary figure S1). Cytokine profiling performed during a symptomatic episode revealed a high serum tumour necrosis factor-α level (56 pg/mL; reference <4 pg/mL) and a high serum interleukin-6 level (11.3 pg/mL; reference <4 pg/mL), although no obvious focus of infection was noted. She had hypolacrima and hyperhidrosis and was diagnosed as having dysautonomia. Motor development of the patient was only slightly delayed. At 8 years of age, she suffered from a severe infection accompanied by coma. Rapid ACTH test indicated adrenal insufficiency. Triple A syndrome was suspected and a leucocyte DNA sample was obtained. She had no mutation in AAAS, and the syndrome was excluded. Despite the feeding problems, she showed catch-up growth reaching to the 50th percentile of height and weight (figure 1B). However, from age 8 years to 11 years, she was treated with high doses of hydrocortisone (20–30 mg/m2/day) and received repeated pulses of prednisolone, and her growth was stunted during that period. The patient died at age 12 years from multiorgan failure due to severe infection. Throughout the course, the patient did not have significant thrombocytopenia, anaemia or any other haematological features. She had a 46,XX karyotype and had no genital abnormality.
Mutation detection
Written informed consent was obtained from the parents of the patients. Leucocyte genomic DNA of patient 1 (obtained at age 4 years) was subjected to long-range PCR to amplify the single coding exon of SAMD9, which was followed by analysis with Nextera XT DNA Library Preparation Kit and MiSeq (Illumina). The identified mutations were verified by Sanger sequencing. Subcloning of the PCR product was performed with TOPO TA Cloning Kit (Thermo Fisher Scientific). Hair follicle-derived genomic DNA was obtained at age 10 years and sequenced. As for patient 2, whole exome sequencing on HiSeq2000 (Illumina) with exome enrichment using NimbleGen SeqCap EZ Human Exome Library V.2.0 was performed. The identified mutations were verified by Sanger sequencing. Long-range PCR with use of WT sequence-specific primers (5′-CAC AGG GAA ATT TTG ACT GAA C-3′ and 5′-CTG TTC TCC AAT TTC AGA AAA AT-3′) and mutant sequence-specific primers (p.Gln39* 5′-CAC AGG GAA ATT TTG ACT GAA T-3′ and p.Asp769Gly 5′-CTG TTC TCC AAT TTC AGA AAA AC-3′) was conducted.
X chromosome inactivation assay
For patient 2, X chromosome inactivation pattern was analysed as previously reported.6 In brief, PCR to amplify polymorphic genomic region of X chromosome was performed using genomic DNA treated with a methylation-sensitive HpaII restriction enzyme. Obtained fragments were analysed with an ABI3500xl sequencer and GeneMapper V.3.7 (Thermo Fisher Scientific).
Functional analyses
Cell lines that express each SAMD9 protein (WT, Ala722Glu or Asp769Gly) in the presence of an inducer, doxycycline, were established. We modified the pBQM812A-1 vector (System Biosciences) with replacing the cumate-inducible promoter by a tetracycline-inducible promoter and the CymR repressor sequence by the reverse tetracycline transactivator sequence. Human SAMD9 cDNA was inserted using the Gibson assembly technique. Two mutations were introduced with a standard site-directed mutagenesis method. Inducible stable HEK293 cells, established according to the protocol of pBQM812A-1, were seeded into 96-well plate with at about 5% confluence in 100 µL of culture medium with or without 1 µg/mL doxycycline. The degree of confluence was quantified every 3 hours using an IncuCyte ZOOM time-lapse microscope (Essen BioScience). Growth curve data are representative of three independent assays that were performed in triplicate.
Results
Mutation detection
In the leucocyte DNA of each patient, two novel heterozygous SAMD9 mutations were detected (figure 2A). The four mutations were not observed in the parents (figure 2B) and were absent in the variant databases (dbSNP138, 1000 Genomes and gnomAD). The two missense mutations (p.Ala722Glu and p.Asp769Gly) occurred in evolutionarily conserved residues (see online supplementary figure S2) and were located in the P-loop NTPase domain (figure 2C).
Molecular findings of the two patients. (A) Partial sequence electropherograms of PCR-amplified SAMD9 are shown. In leucocyte DNA samples of the two patients, two heterozygous mutations (one missense and one nonsense) were detected (indicated by arrows). In the DNA isolated from hair follicles of patient 1, only the missense mutation was detected (indicated by an arrowhead). (B) Pedigrees of patient 1 and patient 2 are shown with SAMD9 genotype information. Black symbols indicate individuals affected by the myelodysplasia, infection, restriction of growth, adrenal hypoplasia, genital phenotypes and enteropathy syndrome. The mutations were only found in these individuals. NA denotes genotype not available. (C) The structure of the SAMD9 protein: the domain architecture was based on Mekhedov et al. 10 The upper lines indicate the locations of previously reported activating mutations, while lower lines indicate mutations identified in the present study. Note that about half of the mutations, and the ones found in this study, are located in the P-loop NTPase domain. (D) Colocalisation of the two mutations on the same SAMD9 allele was tested by a subcloning technique (left panel; patient 1) or by sequence-specific PCR amplification (right panel, patient 2). The results indicate that the two mutations are located on the same allele. (E) X chromosome inactivation analysis using a methylation-sensitive restriction enzyme HpaII. In patient 2, enzyme digestion of genomic DNA resulted in disappearance of one fragment (indicated by an arrow), suggesting non-random methylation of the X chromosome (ie, skewed X chromosome inactivation). (F) Functional effects of the Ala722Glu-SAMD9 and Asp769Gly-SAMD9 were tested with expression experiments. Induced expression of the wild type SAMD9 protein (black circles) caused slower cell growth than non-induced HEK293 cells (white circles). Contrastingly, expression of mutant SAMD9 proteins (Ala722Glu, yellow circles; Asp769Gly cyan circles) resulted in profound growth repression, indicating that they are indeed activating mutations. Bars indicate SEs.
Patient 1 had c.2053C>T, p.Arg685* (relative mutation abundance (RMA) 46%, read depth 42154×) and c.2165C>A, p.Ala722Glu (RMA 51%, read depth 33164×). Subcloning analysis confirmed the colocalisation of the two mutations in the same allele (figure 2D). In hair follicle-derived DNA, only p.Ala722Glu was detected (figure 2A). Patient 2 had c.115C>T, p.Gln39* (RMA 52%, read depth 141×) and c.2306A>G, p.Asp769Gly (RMA 45%, read depth 214×) (figure 2A). PCR with sequence-specific primers was performed (see online supplementary figure S3), showing that the WT/WT primer pair and the mutant/mutant primer pair gave rise to PCR amplification (figure 2D).
X chromosome inactivation assay
Analysis of PCR products that were amplified from genomic DNA digested with methylation-sensitive restriction enzyme showed a uniform X-chromosome inactivation pattern in leucocytes of patient 2 (figure 2E), indicating non-random methylation of the X chromosome.
Functional analyses
To evaluate the effect of the identified missense mutations (p.Ala722Glu and p.Asp769Gly) on cell growth, we expressed each SAMD9 protein (WT or mutant) in HEK293 cells and assessed the cell growth. As previously reported,1 3 induced expression of the WT SAMD9 protein caused mild growth suppression (figure 2F). Contrastingly, expression of the mutant proteins caused severe growth suppression (figure 2F), suggesting that these two mutations excessively activate the intrinsic growth-restricting capacity of the WT SAMD9 protein (ie, activating mutations).
Discussion
We describe two patients with MIRAGE syndrome with two de novo heterozygous SAMD9 mutations on the same allele. For patient 1, p.Arg685* was identified as a somatic mutation. Considering the potent growth-restricting capacity of p.Ala722Glu-SAMD9, additional acquisition of the nonsense mutation would revert the growth restriction (ie, second-site reversion mutation). The heterozygous state of p.Arg685* suggests that the haematopoietic cell pool would be completely replaced by the revertants. For patient 2, although we could not test whether p.Gln39* was a somatic mutation since the patient has already died, skewed X chromosome inactivation in leucocytes was shown, indicating their monoclonality. Simultaneous detection of one activating and another inactivating SAMD9 mutation in MIRAGE syndrome have already been reported by Buonocore et al,3 although colocalisation of two mutations on the same allele was not confirmed. In the report by Buonocore et al, two out of four patients with two mutations died very early, and the remaining two patients developed myelodysplastic syndrome with monosomy 7.3 Probably due to these reasons, the proportion of the revertant cells did not reach to 100%.3 Hence, the two patients described in this study are the first patients who achieved complete haematological reversion. Such haematological reversion has been repeatedly observed in patients with Fanconi anaemia7 and dyskeratosis congenita.8 Very recently, haematological reversion was also observed in one patient with ataxia-pancytopenia syndrome due to an activating SAMD9L mutation.9
Thrombocytopenia and/or anaemias in early infancy have been the most consistent manifestation of the MIRAGE syndrome, seen in 16 out of 17 evaluated patients.1 3 5 This fact highlights the significance of our two patients lacking haematological features. One possible explanation is that the expressivity of the haematological phenotype is variable and it might be absent with certain probability. More plausible explanation, which we favour, considers the effect of the second-site reversion mutations. If the haematopoietic reversion occurred in utero, haematological phenotype could be avoided or alleviated. To reach firm conclusions, studies using serial DNA sampling, including the neonatal period, will be necessary.
There are several additional features that are probably associated with the syndrome. Patient 2 showed clear dysautonomic symptoms, including hypolacrima, hyperhidrosis and blood pressure dysregulation. Renal manifestations (interstitial nephritis and renal tubular defects) are potential late complications of the long-lived patients. The SAMD9 protein is expressed in the glomeruli and renal tubules (Human Protein Atlas; http://www.proteinatlas.org/), and it is possible that SAMD9 mutations affect these tissues. Patient 2 showed remarkable catch-up growth. Most patients with MIRAGE syndrome have feeding problems due to lesions of the oesophagus and/or colon. These situations make it difficult to discriminate whether growth restriction is due to the intrinsic nature of the syndrome or due to suboptimal nutrition. The growth pattern of patient 2 indicates that at least a subset of patients can achieve normal postnatal growth.
To conclude, we described two patients with MIRAGE syndrome who had an activating SAMD9 mutation, and a second-site reversion nonsense mutation in the haematopoietic cells. The phenotype of the patients were characterised by the lack of haematological features. Our experiences expand the knowledge about the phenotypic spectrum of the syndrome.
Acknowledgments
The authors thank Erina Suzuki, Saori Miyasako and Dana Landgraf for their technical assistance.
Footnotes
HS and KK contributed equally.
Contributors YN, KS, AS and KM clinically characterised the patients. TO, MF and KK collected genetic samples. HS, KK, RJ and MS conducted sequencing experiments. HS conducted expression experiments. KK and SN created the design of the study. HS and SN wrote the manuscript with critical inputs from KK, AH and MS. All authors read and approved the final manuscript.
Funding This work was partly supported by a DFG grant HU 895/5-2 (Clinical Research Unit 252), a DFG grant KO 3588/2-1, Takeda Science Foundation, Japan Intractable Diseases (Nanbyo) Research Foundation, the Grant-in-Aid for Scientific Research on Innovative Areas from MEXT (3905-A02) and a grant from the Practical Research Project for Rare/Intractable Diseases of the Japan Agency for Medical Research and Development, AMED.
Competing interests None declared.
Patient consent Obtained.
Ethics approval National Center for Child Health and Development, Medical Faculty, Technische Universität Dresden.
Provenance and peer review Not commissioned; externally peer reviewed.