Kids First-KOMP2: Precision Modeling of Pediatric Conditions Pilot 

The NIH Gabriella Miller Kids First Pediatric Research Program (Kids First) and the Knockout Mouse Phenotyping Program (KOMP2) are collaborating on a pilot project to develop mouse strains to study, phenotype, and validate coding and noncoding genetic variants (e.g. missense, structural variants, copy number variants, INDELS, frame shifts) identified from Kids First datasets. 

The Kids First and KOMP2 leadership teams hosted a webinar titled “Precision Modeling of Pediatric Conditions” on September 21, 2018 to provide information about this collaboration. Investigators can watch the webinar and review PowerPoint slides from the presentations for details about this opportunity. 

To learn more about KOMP2’s pipeline visit, www.mousephenotype.org. For technical questions related this Kids First-KOMP2 collaboration contact: KidsFirstKOMP@nih.gov. Updates about this project will be announced once information become available. 

Projects Selected for this Opportunity: 
 

• Analysis of variation in a cohort of 100 patients with PHACE syndrome
  Investigation Team: Nara Sobreira & Elizabeth Partan, Johns Hopkins University; Dawn H. Siegel & Monika Tutaj, Medical College of Wisconsin
  Project Narrative: PHACE association is a rare disorder defined as the presence of a large segmental hemangioma (usually facial) plus one or more structural brain, cerebrovascular, cardiovascular, ocular, ventral, and occasionally endocrine anomalies. No familial cases have been described, and no causative genes or pathways have been definitively identified. As part of the Gabriella Miller Kids First Research Program, we performed WGS on a cohort of patients with PHACE association and identified rare, de novo, coding and noncoding single nucleotide variants and indels. We prioritized candidate causative variants if: 1) they were not present in gnomAD v2 and v3; 2) they were classified as pathogenic or likely pathogenic in ClinVar; 3) they affected a gene already known to cause a phenotype overlapping that of PHACE; 4) they affected a gene where the mouse model phenotype overlapped that of PHACE; 5) the gene was mutated in multiple probands; or 6) the gene was found to be affected in CNVs from multiple patients with facial hemangioma from the database DECIPHER. This analysis identified 23 variants in 12 genes in 20 patients, including coding variants in RASA3 and THBS2. To assay the functional pathogenicity of these variants, we are using CRISPR/Cas9 to knock the variants in to an endothelial cell line. We are also partnering with The Centre for Phenogenomics in the Kids First-KOMP2 collaboration who will use Cas9 to produce the variants in mouse models. We anticipate these analyses will provide a better understanding of the molecular pathogenesis of PHACE association and verify the first causative genetic variants.
  Kids First Dataset Used in This Project: Genomic Analysis of a Cohort with Infantile Hemangiomas Associated with Multi-organ Structural Birth Defects

• Precision Modeling of Adolescent Idiopathic Scoliosis: SSPO and HAPLN1
  Investigation team: Jonathan Rios and Carol Wise, Texas Scottish Rite Hospital and University of Texas Southwestern Medical Center
  Project Narrative for SSPO: Scoliosis is a rotational spinal deformity that typically presents in otherwise healthy, adolescent children. The prevalence of adolescent idiopathic scoliosis (AIS) is 3%, with over 29 million cases reported worldwide. Although a few loci have been associated with disease susceptibility, most of the genetic contributions to AIS await discovery.  To discover new AIS disease mutations we partnered with the Gabriella Miller Kids First (Kids First) initiative to perform whole-genome sequencing of multi-generational families with inherited AIS. This identified rare missense variants in the SSPO gene, encoding subcommissural organ spondin, or “sco-spondin”, segregating with AIS in five independent families. Sco-spondin is a large secreted member of the thrombospondin protein family that is necessary for proper development of the pineal gland and posterior commissure.  During early neurogenic development sco-spondin is secreted into the cerebrospinal fluid, forming a long extracellular thread called the Reissner fiber. Recently, mutations in zebrafish sspo that caused loss of the Reissner fiber in the central canal and third brain ventricle were associated with body axis deformity. Independently, our collaborative team has identified a functional mutation in zebrafish sco-spondin that yields a scoliosis-like phenotype.  In partnership with the Kids First-KOMP2 project, The Centre for Phenogenomics (TCP) team will use CRISPR/Cas9 editing to develop a mouse line with a Sspo missense mutation that we identified in a large family with multiple cases of severe AIS. From conservation constraint and protein-level measures this particular mutation is predicted to be deleterious to protein function. Together with the TCP team we will develop a comprehensive phenotyping strategy to assess the effects of the Sspo mutation. We expect that these approaches will yield a novel model of AIS that will be useful for further mechanistic and translational studies.
  Project Narrative for HAPLN1: Scoliosis is a rotational spinal deformity that typically presents in otherwise healthy, adolescent children. The prevalence of adolescent idiopathic scoliosis (AIS) is 3%, with over 29 million cases reported worldwide. Common variants in GPR126, PAX1, and other genes has been associated with human AIS. Pathway analysis of human AIS GWAS as well as studies of model organisms suggest convergence on Sox signaling that regulates cartilage biogenesis, specifically vertebral cartilage/intervertebral disc. However, most of the genetic contributions to AIS await discovery.  As part of the Gabriella Miller Kids First (Kids First) Pediatric Research Project on Adolescent Idiopathic Scoliosis (AIS), we performed whole-genome sequencing of multi-generational families with inherited AIS. In two of the families we identified rare missense variants in HAPLN1, encoding Cartilage Link Protein 1 (CRTL1) that co-segregated with AIS and were predicted to be deleterious by molecular modeling. HAPLN1 encodes cartilage link protein 1 (CRTL1), an essential component of the cartilage extracellular matrix, where it links the large proteoglycan aggrecan to hyaluronic acid (HA). These large aggregates stabilize and hydrate the tissue, giving it the ability to resist force and deformation. This role is particularly critical in the cartilage that comprises the intervertebral disc. HAPLN1 is regulated by the SOX5/6/9 trio of transcription factors, similar to GPR126, an adhesion class G protein receptor that is associated with human AIS and results in scoliosis when conditionally knocked out in mice. Using CRISPR/Cas9 editing we have generated a mouse line harboring one of the Hapln1 missense mutations as well as a deletion mutant. In partnership with the Kids First-KOMP2 project, the UC Davis laboratory team will use CRISPR/Cas9 editing to develop a mouse line with the second Hapln1 missense mutation, as well as to generate a conditional-ready Hapln1 strain. As we hypothesize our AIS-associated alleles are loss-of-function, it may be necessary to intercross the Hapln1 mouse lines described here with the conditional allele to study CRTL1 in a tissue-specific manner. Together with the UC Davis team we will develop a comprehensive phenotyping strategy to assess these strains. We expect that these approaches will yield a novel model of AIS that will be useful for further mechanistic and translational studies.
  Kids First Dataset Used in These Projects: Pediatric Research Project on Adolescent Idiopathic Scoliosis (dbGaP accession number: phs001410) 

• Phenotypic characterization of novel candidate risk genes of congenital diaphragmatic hernia using in mice models
  Investigation Team: Wendy K Chung and Yufeng Shen, Columbia University 
  Project Narrative: Congenital diaphragmatic hernia (CDH) is a severe birth defect affecting about 1 in 3000 live births. Previous studies have found CDH is genetically heterogeneous. Our team have analyzed over 700 CDH patient-parents trios with whole genome sequencing data generated by the Gabriella Miller Kids First Pediatric Research Program (Kids First). We identified candidate risk genes based on deleterious de novo variants (single nucleotide variants, short insertions and deletions, and copy number variants), and then prioritized candidate genes by integrating statistical evidence with gene expression patterns during embryonic development. In this collaboration with the Knockout Mouse Phenotyping Program (KOMP2) and Kids First, we aim to functionally characterize the phenotypes of selected novel CDH candidate genes (RYBP, CDC42BPB, AGO1, and CNOT1) in mouse models. Dr. Steve Murray and his team at The Jackson Laboratory will create the mouse models and perform detailed phenotyping. We anticipate this collaboration will functionally validate some of the candidate genes and expand our understanding of disease etiology of CDH and associated congenital anomalies.
  Kids First Dataset Used in This Project: Pediatric Research Project on the Genomic Analysis of Congenital Diaphragmatic Hernia (dbGaP accession number: phs001110) 

• Modeling Enchondroma Development in Mice: Testing the Role of KDM4C in Ollier Disease
  Investigation Team: Nara Sobreira (PI) & Sarah Robbins, Johns Hopkins University
  Project Narrative: Ollier disease is a rare disorder characterized by the development of multiple enchondromas. Patients with Ollier disease are also at increased risk for cancer, mainly chondrosarcoma, brain, and gonadal cancers. Though somatic variants in IDH1 and IDH2 are associated with Ollier disease, to date, no germline variants have been described. As part of the Gabriella Miller Kids First Research Program, we performed WGS on a cohort of patients with Ollier disease and identified rare, heterozygous, germline variants in genes in the HIF-1 pathway in 33% of the patients. One patient had compound heterozygous variants in KDM4C, each inherited from one of his parents. By immunoblot analysis of the patient’s fibroblast, we have established that these variants lead to loss of function at normoxia and hypoxia. To prove that these variants are responsible for tumorigenesis in this patient, we are partnering with the Kids First-KOMP2 project. Using CRISPR/Cas9-gene editing, the Jackson laboratory team will develop a compound heterozygous mouse model and a homozygous mouse model for each of the variants identified in the patient with Ollier disease. Together, we will develop a rigorous anatomical phenotyping protocol that will investigate the enchondroma and cancer development in these mouse models. We anticipate that these mouse models will enable future studies on the pathophysiology of enchondroma formation and on possible pharmacological treatments for patients with Ollier disease.  
  Kids First Dataset Used in This Project: Genome-wide Sequencing to Identify the Genes Responsible for Enchondromatoses and Related Malignant Tumors  (data for this project are not yet available for access)
   
• In vivo modeling of a Novel KIF21A missense variant that causes a syndromic CFEOM TUBB3 phenotype.
  Investigation Team: Elizabeth Engle (PI), Boston Children's Hospital
  Project Narrative:  Specific recurrent missense mutations in KIF21A cause isolated congenital fibrosis of the extraocular muscles (CFEOM), a congenital cranial dysinnervation disorder (CCDD) in which the oculomotor nerve axons are misguided and fail to correctly innervate their target extraocular muscles.  These mutations attenuate KIF21A autoinhibition in vitro, resulting in increased association of this kinesin motor protein with the microtubule cytoskeleton. Microtubules are composed of polymers of alpha- and beta-tubulin isotypes, and CFEOM can also result from specific recurrent missense mutations in the gene that encodes the TUBB3 tubulin isotype. In contrast to KIF21A mutations that cause isolated CFEOM, specific TUBB3 mutations cause CFEOM accompanied by maldevelopment of the brain and additional cranial nerves, as well as a progressive axonal peripheral neuropathy. Moreover, the TUBB3 mutations stabilize microtubules in vitro and reduced the association of KIF21A with the microtubule cytoskeleton. To date, the introduction into mouse of the human KIF21A and TUBB3 missense variants have recapitulated the phenotypes, permitting detailed mechanistic studies. Despite these studies, a definitive mechanistic link between KIF21A and TUBB3 CFEOM has not been defined. Through the Gabriella Miller Kids First Pediatric Research Program (Kids First), we have identified a novel KIF21A missense variant that causes a syndromic CFEOM TUBB3 phenotype rather than isolated CFEOM. We have teamed up with Kids First and the KOMP2 Jackson laboratory team to develop and phenotype this knock-in mouse model. Steve Murray and his team at JAX will perform detailed anatomical and behavioral phenotyping, which will be followed by targeted nervous system phenotyping and mechanistic studies in the Engle lab. We believe that modeling and further studying this Kif21a substitution in mouse will not only confirm variant pathogenicity but will also help us to elucidate a common mechanism through which variants in KIF21A and TUBB3 cause human developmental disorders.
  Kids First Dataset Used in This Project: GMKF: Kids First Pediatric Research Program on Congenital Cranial Dysinnervation Disorders and Related Birth Defects (dbGaP accession number: phs001247)
   
• Deep Phenotypic Characterization of a Noncoding Duplication Underlying Congenital Facial Weakness 
  Investigation Team: Elizabeth Engle (PI) & Arthur Lee, Boston Children's Hospital
  Project Narrative: The Engle lab studies the genetic basis of congenital cranial dysinnervation disorders (CCDDs) which are a set of neurodevelopmental diseases that can affect facial and eye movements. The Engle lab has collected, sequenced and analyzed over 899 whole genome sequences through the Gabriella Miller Kids First research program (Kids First). As a next step in this partnership, the Engle Lab will collaborate with the the Kids First and the Knockout Mouse Phenotyping Program (KOMP2) to study the impact of structural variation on CCDDs, using the mouse as a model system. Together we plan to generate mutant mice bearing a 56 kb tandem duplication of a noncoding region that segregates in a large CCDD family and perform behavioral, anatomical, and molecular phenotyping using state-of-the-art technologies.
  Kids First Dataset Used in This Project: GMKF: Kids First Pediatric Research Program on Congenital Cranial Dysinnervation Disorders and Related Birth Defects (dbGaP accession number: phs001247)
   
• In vivo Modeling for a Novel Tubulinopathy 
  Investigation Team: Elizabeth Engle (PI) & Julie Jurgens, Boston Children's Hospital
  Project Narrative: Ocular congenital cranial dysinnervation disorders (CCDDs) are rare pediatric disorders of neurodevelopment defined by defective innervation of the extraocular muscles, which are responsible for eye movement. Although ocular CCDDs are often inherited, the genetic basis for many of them remains unknown. Through the Gabriella Miller Kids First Pediatric Research Program, the Engle lab has performed whole genome sequencing of 899 individuals with unsolved CCDDs and their families, enabling identification of a novel candidate missense variant in a tubulin isotype in one of these disorders. Tubulin isotypes have high homology to one another and modeling missense tubulin variants in mouse can be challenging. Thus, the Engle lab is teaming up with Kids First and the Knockout Mouse Phenotyping Program (KOMP2) to develop and phenotype a knock-in mouse model. Dr. Kent Lloyd’s team at UC Davis is generating this mouse model and will perform detailed anatomical and behavioral phenotyping, which will be followed by targeted phenotyping of cranial nerve development in the Engle lab. We anticipate that this collaboration will validate a strong novel candidate genetic variant and expand our fundamental understanding of mechanisms underlying ocular CCDDs and other tubulinopathies, enabling improved genetic counseling and care for patients with these conditions. 
  Kids First Dataset Used In This Project: GMKF: Kids First Pediatric Research Program on Congenital Cranial Dysinnervation Disorders and Related Birth Defects (dbGaP accession number: phs001247)
   
• Development of Mouse with 343bp Deletion in Gata4 Intron 2 to Test Role in CDH and CHD
  Investigation Team: Gabrielle Kardon (PI) & Eric Bogenschutz, University of Utah
  Project Narrative: Congenital Diaphragmatic Hernias (CDHs) are common and often lethal birth defects in which an aberrantly developed diaphragm allows abdominal contents to herniate into the thoracic cavity. Using whole genome sequencing of CDH children and their healthy parents, we have discovered an inherited noncoding variant, in which a 350 base pair deletion disrupts a putative enhancer of Gata4, a gene highly associated in humans with CDH and functionally demonstrated in mouse to cause CDH. We hypothesize that this deletion leads to reduced Gata4 expression, a highly dosage sensitive gene, and creates a sensitized genetic background for development of CDH. To test this, we will generate mice with this 350 base pair deletion and test whether this creates a sensitized genetic background for development of CDH. If true, this will provide the first evidence that noncoding genetic elements contribute to the etiology of CDH. We will generate the relevant mutant mice in collaboration with the KOMP2 DTCC Consortium at The Centre for Phenogenomics using CRISPR/Cas9-endonuclease technologies.
  Kids First Dataset Used in This Project: Kids First: Pediatric Research Project on the Genomic Analysis of Congenital Diaphragmatic Hernia (dbGaP accession number: phs001110)

• Noncoding De Novo Variants in Congenital Heart Disease: From Predicted Pathogenicity to Model Organism
  Investigation Team: Bruce Gelb (PI) & Felix Richter, Icahn School of Medicine at Mount Sinai on behalf of the Pediatric Cardiac Genomics Consortium
  Project Narrative: Congenital heart disease (CHD) represents a growing disease burden and is primarily genetic. CHD is typically sporadic with approximately 8% explained by damaging coding de novo variants. Building on these data, we compared noncoding de novo variants in 800 parent/affected-child trios (750 from Gabriella Miller Kids First Pediatric Research Program) to controls using variant-level scores that predict cardiac regulatory disruption. We discovered a burden of predicted damaging variants in cases, including a noncoding de novo deletion in BCOR and a noncoding de novo single base change in GATA4. A critical next step to establish pathogenicity is testing if these mutations cause CHD in mice. To model these variants, we are working with experts in the latest CRISPR/cas9 technology through the Kids First-KOMP2 collaboration. Kent Lloyd’s group from University of California, Davis will create mice with the BCOR deletion, and Lauryl Nutter’s team from The Center for Phenogenomics at Toronto’s Hospital for Sick Children will introduce the GATA4 variant into mouse germlines. Further leveraging the Kids First-KOMP2 partnership, mutant mice will be phenotyped through extensive imaging and pathology protocols. This work is crucial to improving our understanding of how noncoding regions, which account for 99% of the human genome, contribute to birth defects like CHD.
  Kids First Dataset Used in This Project: Discovery of De Novo and Inherited Mutations that Cause Prevalent Birth Defects and Discovery of the Genetic Basis of Structural Heart and Other Birth Defects (dbGaP accession number: phs001138)