The Endocrine Society's 97th Annual Meeting & Expo:
Genomic Technology and Idiopathic Short Stature
New genetic tools for the evaluation of children with short stature were the topic of a talk by Dennis Jay Chia, MD, Assistant Professor, Albert Einstein College of Medicine, New York, New York. Dr. Chia spoke on March 6, 2014, at the Endocrine Society Annual Meeting in San Diego, California.
Available Genetic Tests
- Chromosome analysis
- Testing for specific genes or syndromes, such as SHOX and PTPN11, etc.; Russell-Silver syndrome
At the Interface of Research and Clinical Practice
- Copy number variation (CNV)
- Whole exome sequencing
Copy Number Variants
Rare CNVs have been identified as a common cause of short stature. In an article by Zahnleiter et al (PLoS Genet) in 2013, the yield of pathogenic CNV was estimated to be approximately 10% for short stature. The investigators compared 200 subjects with idiopathic short stature (standard deviation for height, below -2; mean, -2.75) with 820 controls of normal stature. Criteria for defining pathogenic CNV were:
- Exclusion of CNVs <50 kilobase pairs, with significant overlap to those identified in the control population, mapped to intronic or intergenic region.
- CNV segregates with short stature phenotype if familial, de novo if sporadic.
- Gene linked to growth in Online Mendelian Inheritance in Man, murine knockout models, and other potential mechanisms.
Ten deletions/10 duplications (20 pathogenic CNVs) were identified in 200 subjects, for a total of 10% yield. An algorithm for the routine use of CNV analysis was reviewed in 2012 (Mefford et al, N Engl J Med) for the identification of the cause of developmental delay/autism, based upon a yield in the 12-14% range. Therefore, based on a similar yield, one could argue for routine use of CNV analysis in idiopathic short stature.
Whole Exome Sequencing
Whole exome sequencing can also identify underlying mutations for clinical phenotypes. An example from the literature demonstrating the diagnostic efficacy from the lab of Baylor was published in 2013 (Yang et al, N Engl J Med).
Genome-Wide Association Study (GWAS)
Most GWAS SNP variants map to gene regulatory regions, which holds true for GWAS SNPs of individual diseases, such as prostate and breast cancer, systemic lupus erythematosus and Crohn’s disease. Common genetic variants of typically small effect size are found in much greater frequency in regulatory regions than in exons. We would hypothesize that rare genetic variants of large effect size also frequently map to regulatory regions.
Mutations in a Distal Enhancer of PTF1A
There are few examples of mutations of enhancers as the molecular mechanism for a disease presentation. To highlight one recent example, mutations in a distal enhancer of PTF1A were found to be the most common cause of isolated pancreatic agenesis in 2014 (Weedon et al, Nat Genet).
Epigenetic modifications mark gene regulatory regions: chromatin accessibility and DNA methylation
Noncoding genetic variants of regulator elements may be accompanied by local chromatin changes. In fact, chromatin accessibility is a hallmark of domains of regulatory DNA.
Approximately 60 to 80% of cytosine residues of CpG dinucleotides of the human genome are methylated. Classically, CpG islands are identified adjacent to promoters. Methylation is generally associated with transcriptional silencing. There is evolving evidence for local DNA hypomethylation that is associated with other active chromatin features in 2013 (Hon et al, Nat Genet). Altered DNA methylation and imprinting of IGF21/H19 were demonstrated to occur in syndromes of disordered fetal growth in 2005 (Gicquel et al, Nat Genet). Tissue-specific differentially methylated regions were shown to map to distal regulatory elements in 2013 (Hon et al, Nat Genet). Common disease-associated variation frequently maps to cell/tissue-specific regulatory DNA.
Evaluating for epigenetic changes at key loci in growth
Growth hormone Stat5b binds at multiple domains at the Igf1 locus. Recent studies have revealed that Stat5-binding domains of the Igf1 gene are open only in the liver, and that liver-specific DNA hypomethylation at Igf1 enhancer domains mirrors chromatin accessibility in the liver, kidney, and spleen. This illustrates that having the appropriate tissue will likely be necessary to assess for pathogenic epigenetic changes.
Epigenetic silencing of Pdx1 expression has been demonstrated in a rat model of intrauterine growth retardation in 2008 (Park et al, J Clin Invest), and diminished methylation of insulin-like growth factor 2 differentially methylated regions in individuals exposed to famine in early gestation has also been observed (Heijmans et al, Proc Natl Acad Sci USA, 2008). These provide evidence that epigenetic changes can be demonstrated to underlie disease phenotypes. A potential tool to overcome the tissue accessibility barrier, at least in the setting of an underlying genetic cause, is induced pluripotent stem cell disease modeling.
“I wanted to bring attention to currently available clinical testing and research-level testing that may improve our diagnostic yield,” concluded Dr. Chia. Copy number variation and whole exome sequencing are available tools that can identify a molecular cause of short stature in a subset of subjects, but how we incorporate them into clinical practice remains to be seen. Mutations of gene regulatory elements and epigenetic modifications are additional plausible molecular mechanisms. “I reviewed the use of DNA methylation studies to look for functional changes of gene regulatory regions,” Dr. Chia said, “the caveat being that there may likely be tissue-specific changes that represent a barrier in clinical settings.”
March 13, 2015