A New Homozygous IGF1R Variant Defines a Clinically Recognizable Incomplete Dominant form of SHORT Syndrome

Insulin-like growth factors I (IGF1; MIM #147440) and II (IGF2; MIM #147470) act by binding to a common receptor, the insulin growth factor-1 receptor (IGF1R; MIM #147370), which shares a high level of homology with the insulin receptor (INSR; MIM #147670) [Abbott et al., 1992; LeRoith et al., 1995]. Ligand-receptor binding leads to IGF1R receptor autophosphorylation and tyrosine phosphorylation of multiple signaling adapter substrates, including the insulin-receptor substrates (IRS; MIM #147545 and IRS2; MIM #600797), Shc, and 14-3-3 proteins [Jones and Clemmons, 1995]. In turn, substrate phosphorylation activates two main signaling pathways, the PI3K/AKT/mTOR and the Ras-MAPK.

SHORT syndrome (MIM #269880) is a genetically heterogeneous autosomal-dominant condition caused by heterozygous mutations in PIK3R1 (MIM #171833) encoding for the regulatory subunit (p85alpha) of the phosphatidylinositol 3 kinase (PI3K). Patients with SHORT syndrome show intrauterine growth retardation (IUGR), short stature, hyperextensibility of joints and/or hernias, ocular depression, Rieger anomaly, delay of tooth eruption. Furthermore, they have a recognizable facial gestalt characterized by progeroid and triangular facies, lack of facial fat, and hypoplastic alae nasi. Partial lipodistrophy, insulin resistance, nephrocalcinosis, hearing deficits, atrial and/or ventricular septal defects, and disorder of lipid metabolism are additional features of SHORT syndrome. Developmental milestones and cognition are normal. In vivo studies using patient's fibroblasts suggested that PI3KR1 mutations impair the PI3K-AKT-mTOR pathway [Chudasama et al., 2013; Dyment et al., 2013; Thauvin-Robinet et al., 2013].

Heterozygous mutations in the IGF1R gene, causing partial resistance to IGFI, have been detected in patients with nonsyndromic IUGR and postnatal growth failure [Kawashima et al., 2005; Wallborn et al., 2010], whereas three patients with compound heterozygosity and one with a homozygous mutation have been described so far [Abuzzahab et al., 2003; Fang et al., 2012; Gannage-Yared et al., 2013]. The latter displays a more severe phenotype mainly characterized by IUGR, short stature, developmental and speech delay, microcephaly, facial dysmorphism, heart malformation, and reduced subcutaneous fat.

Here, using exome sequencing, we identified and characterized a new homozygous IGF1R missense mutation in a child showing clinical features of SHORT syndrome as well as high IGFI levels, CNS defects, developmental delay, and a pronounced progeroid appearance. Proband's parents and grandmothers were heterozygous carriers of the mutation and they presented with high IGFI levels, short stature, and type 2 diabetes. Functional studies using patient's primary skin fibroblast cell cultures revealed a lower IGF1R expression in patient's cells that alters PI3K/AKT/mTOR signaling and autophagy, thus defining the pathological mechanism underlying this newly and clinically recognizable incomplete dominant form of SHORT syndrome.

The patient (Supp. Fig. S1A) was born from consanguineous Italian parents (cousin once removed) (Supp. Fig. S1B) at 36 weeks gestation by caesarean section for a severe IUGR. Her birth weight was 950 g (−4.12 SD), length was 39 cm (−3.20 SD), and OFC was 24.5 cm (−5.53 SD); the APGAR scores were 8 at 1’ and 9 at 5’. She was admitted to the intensive neonatal care for her poor general condition; she ingested up to 120 kcal per kilogram of body weight per day, but due to the absence of linear growth a parenteral nutrition was started.

During the hospitalization in her first month of age, she underwent a wide diagnostic evaluation. She showed anemia and thrombocytosis. The otoacoustic emission was absent and the evoked auditory potential evaluation confirmed the presence of a bilateral sensorineural hearing loss, mainly in the 2,000–4,000 Hz range. The presence of a cardiac murmur prompted an echocardiography showing an interatrial septal defect of 5 mm ostium secundum type, a patent foramen ovale (FO) and two narrowed pulmonary branch arteries. The subsequent echocardiography, performed at 7 months, showed normal interatrial septum and closed FO, but increased Doppler gradient of the right pulmonary arterial branch. The brain MRI showed a hypointense signal of the supratentorial white matter in T1 that was hyperintense in T2, with an associated underrepresentation of the cerebral gyri and a hypo-plastic corpus callosum. The oculistic evaluation disclosed an atrophic left iris. Metabolic screening that included aminoacidemia, urinary acid organic, and ammonemia resulted normal. A total body X-ray was performed coupled with abdominal ultrasound scans that were marked normal. Other standard laboratory analyses were normal, except for triglycerides (189 mg/dl, normal range, [NR] 40–159), calcium (10.6 mg/dl, NR 8–10.5), phosphorus (6.5 mg/dl, NR 3.5-7), and magnesium (2.59 mg/dl, NR 1.71-2.29). Serum IGFI concentration was evaluated twice, at 8 months and at 14 months, resulting 285 and 352 ng/ml, respectively (NR 18–146 ng/ml) [Granada et al., 2000]. The serum levels of growth hormone, insulin, glucose, and Hb1Ac were in the normal range (Table 1).

At the time of our examination, 9 months, her length was 57 cm (−3.20 SD), weight was 3,170 g (−4.0 SD), and OFC was 35 cm (−5.3 SD), and she showed a marked progeroid appearance along with dysmorphic features (Table 1; Supp. Fig. S1A). At 24 months, the neuropsychiatric evaluation determined the presence of a global developmental delay, inability to walk alone, axial hypotonia and mild hypertone in the lower limbs, normal eye contact and behavior, and poor language development.

Her 35-year-old mother was 151 cm (−2.3 SD), her 38-year-old father was 160 cm (−2.2 SD). IGF-I values, analyzed on September 2014, were 350 ng/ml (NR 73–244) in the mother and 263 ng/ml in the father (NR 69–226). The maternal grandfather was 190 cm tall and the paternal grandfather was 165 cm, whereas both the consanguineous grandmothers were 148 cm (−2 SD) and they developed type 2 diabetes, reported also in other relatives (Supp. Fig. S1C). In both grandmothers’ families, several relatives also showed growth retardation and short stature (Supp. Fig. S1C).

We exome-captured and sequenced the affected sibling and their parents. Variants were annotated with snpEff v2.0.2 against Ensembl v75 to identify new damaging mutations. After filtering out the variants recorded, the remaining variants were sorted according to their inheritance pattern and putative pathogenicity (Supp. Table S1). We uncovered a single putatively deleterious variant that segregated with the disease (see Supp. Methods for more details). This variant was homozygote in the probands and heterozygote in the parents and thus was compatible with an autosomal-recessive inheritance of the disease (Supp. Fig. S1C). Notably, the heterozygous variant was found in all family members with milder phenotype including high IGFI levels, short stature, and type 2 diabetes, for whom DNA was available (Supp. Fig. S1C). The identified variant is a novel homozygous IGF1R missense variant, c.2201G>T in NM_ NM_000875.4 (nucleotide numbering uses +1 as the A of the ATG translation initiation codon in the indicated reference, with the initiation codon as codon 1), locates at the last nucleotide of exon 10, a position possibly affecting splicing process as predicted by NetGene2 version 2.42 [Brunak et al., 1991] and NNSPLICE version 0.9 [Reese et al., 1997]. We therefore assessed the transcriptional level effect of this mutation through RT-PCR. Direct sequencing of the amplified products revealed the presence of an aberrant isoform generating a mutant protein containing 25 additional amino acids, p.Pro733_Arg734ins25 (Supp. Fig. S1D).

After patient and parent's skin fibroblasts generation, we therefore investigated the consequence of the c. 2201G>T variant at protein level by assessing the activity of IGF1R pathway by measuring the IGF1R abundance, IGF1R phosphorylation level, and the activation of the IGF1R pathways and autophagy, in cells treated or not with IGFI and/or insulin (see Supp. Methods for more details).

Both patient and heterozygous carrier's cells expressed lower levels of IGF1Rβ subunit, compared with normal cells (Fig. 1A). Immunoblotting against phosphorylated tyrosine residues of IGF1R β-subunit showed diminished autophosphorylation in patient's fibroblasts (Fig. 1A). The binding of IGFI to the IGF1R receptor initiates a signaling cascade of binding and tyrosine phosphorylation of downstream molecules that ultimately leads to the activation of the major signaling pathways such as the protein kinase β/AKT. Our analysis revealed that both patient and parent's fibroblast cell lines showed pSer473-AKT lower levels, which marks AKT activity when compared with a normal individual cells (Fig. 1B).

Next, to examine whether the IGF1R mutation may impact on insulin signaling [Girnita et al., 2014], we exposed the patient, parent, and control fibroblast cell lines to increasing insulin concentration. We found that total insulin receptor was equally expressed in cell lysates from IGF1R mutant fibroblasts and matched controls (Fig. 1C), whereas lower pAKT protein level was found in the patient and to a less extent in parent's fibroblast cell line (Fig. 1D).

Overall, these results demonstrated an impaired capability of the mutated IGF1R to phosphorylate itself and to activate the IGFI or insulin-AKT downstream signaling pathway.

Based on the PI3K/AKT-mediated responsiveness of active mTORC1 in suppressing autophagy, we speculated that a defective IGF1R should result in activation of the autophagy process. To assess this hypothesis, we measured the autophagy flux level in patient's fibroblast cells by monitoring LC3-II and p62/sequestosome-1 protein levels, two of the main autophagy protein markers. We detected an increase of endogenous levels of LC3-II and a parallel decrease of p62 protein levels in patient cells treated with the autophagy inhibitor bafilomycin A1, compared with normal cells (Fig. 1E). Overall, these experiments showed that IGF1R mutation causes an increase of autophagy flux.

Using exome sequencing, we identified a new homozygous IGF1R missense mutation in a patient showing SHORT syndrome plus additional features. We studied the functional effect of this variant both at RNA and protein level, demonstrating that this nucleotide substitution leads to an abnormal splicing event producing a 25 amino acids longer protein (Supp. Fig. S1C) that affects PI3K/AKT/mTOR signaling and autophagy.

To date, four patients have been described harboring biallelic mutations in IGF1R [Abuzzahab et al., 2003; Fang et al., 2012; Gannage-Yared et al., 2013]. Our patient shared many clinical (IUGR, short stature, lipodystrophy, microcephaly, developmental delay), dysmorphic, malformative (narrowed pulmonary branch arteries), and metabolic (high triglyceride, calcium, and phosphorus) features (Table 1) with the reported cases. Interestingly, our case also showed CNS anomalies, ocular defects (Rieger anomalies), deafness, and a more pronounced progeroid appearance, thus positioning this case at the extreme end of the clinical spectrum associated with IGF1R mutations. The presence of shared phenotypes supports the existence of a specific and clinically recognizable syndrome clearly overlapping with the SHORT syndrome (Table 1). Yet, the biochemical evidence that IGF1R and PIK3R1 mutations affect the same PI3K/AKT/mTOR pathway prompt us to suggest the term of “SHORT syndrome type 2” to define this incomplete dominant condition distinguishable from SHORT syndrome by the recessive mode of inheritance, high IGFI level, development delay, intellectual disability, CNS defects, and a more pronounced progeroid appearance. It is conceivable that SHORT syndrome and SHORT syndrome type 2 belong to a more genetically heterogeneous group of “IGF1R/PI3K-signaling disorders,” where mutations in genes acting in the IGF1R signaling pathway may result in overlapping phenotypes.

A differential clinical diagnosis of neonatal progeroid syndrome (NPS), also called Wiedemann-Rautenstrauch syndrome (MIM #264090), was also possible for the index case due to the evident clinical analogies (Table 1). NPS represents a particular form of progeria, with manifestations of prenatal and postnatal growth deficiency, pseudo-hydrocephalus, small and already senile-appearing face since birth, hearing loss, deficiency of adipose tissue, congenital heart, and CNS defects [Toriello, 1990]. NPS etiology is still unknown; however, it is remarkable that experimental evidence showed that somatotroph signaling might be impaired. Indeed, recent studies on a murine model of progeroid laminopathies, the Zmpste24-deficient mice, showed a relevant dysregulation of GH/IGF1 balance, with a progressive reduction of blood IGFI levels, relative increase in GH levels, and profound transcriptional alterations in key genes for somatotroph signaling [Haigis and Yankner, 2010].

These observations suggested that somatotroph axis defects might be detrimental, causing features of progeroid phenotype, such as reduced growth rate and body. Moreover, the treatment of Zmpste24-deficient mice with recombinant IGFI (rIGFI) ameliorated some of the progeroid features, such as improvement of body weight, increased amount of subcutaneous fat deposits, reduction of lordokyphosis, alopecia, and significant extension of life span. For this reason, rIGFI treatment has been recently proposed as a potential therapeutic approach to slow down disease progression in children with progeroid syndrome [Ugalde et al., 2010]. In this regard, our experimental biochemical data in one of the main effector of the somatotroph axis pathway suggest that this dysregulation might be responsible not only for the growth defects and the lipodystrophy, but might be involved also in some of congenital malformation and functional deficits seen in our patient (Table 1). However, animal modeling and in vitro studies need to be further explored in detail.

Furthermore, recent findings have unexpectedly shown that progeroid murine models exhibit an activation of autophagy, instead of the characteristic decline in this pathway occurring during normal aging [Marino et al., 2008]. Autophagy is an evolutionarily conserved process essential for the maintenance of cellular homeostasis and organismal viability regulated by mTOR inhibition and AMPK activation [Mizushima et al., 2008]. In line with these evidence, our results, which show an autophagy increase as a consequence of IGF1R mutation, establishes the presence of common pathogenetic events in patients with SHORT syndrome type 2 and those with progeroid syndromes.

Therefore, a clear distinction of patients with SHORT syndrome type 2 from the ones with NPS, especially at birth, could be challenging. Based on our experimental and clinical evidence, we suggest that newborns showing a suspected NPS phenotype should receive an IGFI evaluation and, if abnormally high, a molecular analysis of IGF1R. This has relevant prognostic and recurrence risk implications, as well as possible therapeutic impact, since the treatment with rIGFI is expected to be less effective in these patients.

In conclusion, we defined a new and clinically recognizable incomplete dominant form of SHORT syndrome, which needs to be taken into account in the 10% of IUGR infants [Abuzzahab et al., 2003], and in the differential diagnosis with the dominant form of SHORT syndrome and with the NPS.

Disclosure statement: The authors declare no conflict of interest.