Somatic mutations and progressive monosomy modify SAMD9-related phenotypes in humans

It is well established that somatic genomic changes can influence phenotypes in cancer, but the role of adaptive changes in developmental disorders is less well understood. Here we have used next-generation sequencing approaches to identify de novo heterozygous mutations in sterile α motif domain–containing protein 9 (SAMD9, located on chromosome 7q21.2) in 8 children with a multisystem disorder termed MIRAGE syndrome that is characterized by intrauterine growth restriction (IUGR) with gonadal, adrenal, and bone marrow failure, predisposition to infections, and high mortality. These mutations result in gain of function of the growth repressor product SAMD9. Progressive loss of mutated SAMD9 through the development of monosomy 7 (–7), deletions of 7q (7q–), and secondary somatic loss-of-function (nonsense and frameshift) mutations in SAMD9 rescued the growth-restricting effects of mutant SAMD9 proteins in bone marrow and was associated with increased length of survival. However, 2 patients with –7 and 7q– developed myelodysplastic syndrome, most likely due to haploinsufficiency of related 7q21.2 genes. Taken together, these findings provide strong evidence that progressive somatic changes can occur in specific tissues and can subsequently modify disease phenotype and influence survival. Such tissue-specific adaptability may be a more common mechanism modifying the expression of human genetic conditions than is currently recognized.

A Tandem Mass Spectrometry assay instead of immunoassay; B post twice weekly stimulation with human chorionic gonadotropin (hCG) for 3 weeks; C basal and peak cortisol in cosyntropin stimulation test.
Normal ranges can be dynamic and influenced by preterm delivery. Basal testosterone typically falls in the first 5 days after delivery and then rises between 50-90 days in the "minipuberty of infancy" to peak around 300-450 ng/l. The timing of the testosterone rise in preterm infants may be delayed compared to term babies. AMH concentration in first 2 weeks of life is typically 30 ng/ml. Cortisol basal values are typically 5-14 µg/dl and peak response to cosyntropin >20 µg/dl. ACTH normal range 10-60 pg/ml.
Note: immunoassays can be unreliable in preterm infants because of cross reactivity with fetal adrenal steroids, although in these children the fetal adrenal output was generally low.  Figure 1. SAMD9 is expressed in human fetal testis and adrenal gland.
(A) Immunohistochemistry showing SAMD9 expression in human fetal testis at 9 weeks post conception (red) with NR5A1 (also known as steroidogenic factor-1) shown in green and DAPI-stained nuclei in blue. Control data without the primary antibody are shown below. The thin layer of DAPI-stained cells is the capsule (C). NR5A1 highlights the nuclei of interstitial Leydig cells (L). SAMD9 appears in the cytoplasm of both interstitial cells and cells in the seminiferous cords (SC). Scale bar, 100μm. (B) Control data for the fetal adrenal immunohistochemistry (Figure 2A). There is some autofluorescence of the cytoplasm of adrenal cells in the green channel, but the pattern is distinct from the nuclear staining of NR5A1. Scale bar, 100μm. Electropherograms of single nucleotide primer extension reactions in genomic DNA are shown from the samples described. PCR products were generated to flank the c.1376G>A or c.2054G>A mutation and single nucleotide extension from a specific 35mer primer was used to probe the nucleotides at these positions. Fragments were separated on the ABI-3730 sequencer. The position of a 35b marker is indicated.  The allele percentages were determined using single-nucleotide primer extension assays. (B) Blood leukocytes from the two patients who had early monosomy 7 and myelodysplasia. Allele percentages were determined by PCR amplification and subcloning of DNA followed by sequencing of more than 50 independent clones.

STOP STOP C A T T A G A T T C A T A A G A T T G G T C G A G C T G G T T G A G C T
Supplemental Figure 5. Low-level nonsense mutations detected in Patients 3 and 5 on deep sequencing were confirmed by subcloning. Mutant clones are shown above (mutant nucleotide underlined) and wild-type clones are shown below.

Mutant
Wild-type

Patient 5 Patient 3
Patient 8 Patient 7 Supplemental Figure 6. Confirmation of additional loss-of-function nonsense and frameshift changes following subcloning of PCR-amplified DNA from Patients 3, 5, 7 and 8. Parental DNA was also subcloned for these regions (Patients 3, 5 and 7) and did not show any mutations. Figure 7. Number of vesicles greater than 1.5 μm diameter per cell. Data are presented for 20 individual fibroblasts from each patient or control studied by electron microscopy. No statistically significant differences were seen between groups. Figure 8. iPS cells generated from Patients 6 & 8 were karyotyped and sequenced in order to confirm no large chromosomal changes had occurred and that the primary SAMD9 mutation was present.