Two novel bi‐allelic KDELR2 missense variants cause osteogenesis imperfecta with neurodevelopmental features

Department of Neuromuscular Disorders, Queen Square Institute of Neurology, University College London, London, UK Section of Pediatric Neurology and Developmental Neuroscience, Department of Pediatrics, Baylor College of Medicine, Houston, Texas Department of Molecular and Human Genetics, Baylor College of Medicine, Houston, Texas Texas Children's Hospital, Houston, Texas Development and Behavioural Pediatrics Department, Institute of Child Health and The Children Hospital, Lahore, Pakistan Division of Neuroradiology, Edward B. Singleton Department of Radiology, Texas Children's Hospital, Houston, Texas Department of Medical Genetics, Health Sciences University, Istanbul Kanuni Sultan Suleyman Research and Training Hospital, Istanbul, Turkey Department of Pediatrics, Faculty of Medicine, Kuwait University, Safat, Kuwait Human Genetics Center, University of Texas Health Science Center at Houston, Texas Human Genome Sequencing Center, Baylor College of Medicine, Houston, Texas Department of Pediatrics, Baylor College of Medicine, Houston, Texas


To the Editor:
In a recent article in the American Journal of Human Genetics, biallelic pathogenic KDELR2 variants were described as a novel cause of autosomal recessive (AR) osteogenesis imperfecta (OI) (MIM: #166200) in four families with six affected individuals (van Dijk et al., 2020). The KDELR family of proteins is important in interorganelle communication by regulating protein trafficking between the Golgi apparatus and the endoplasmic reticulum (Capitani & Sallese, 2009). KDELR2-related OI results from the inability of HSP47 (heat shock protein 47) to bind KDELR2, leading to failure of HSP47 to dissociate from collagen type 1. HSP47-bound extracellular collagen cannot form collagen fibers in individuals with pathogenic biallelic KDELR2 variants (Figure 1; van Dijk et al., 2020). We read the authors' work with great enthusiasm and would like to share clinical and genetic information from two additional unrelated consanguineous families with three affected children with OI with additional phenotypic features, therefore expanding the phenotypic spectrum of KDELR2-related OI.
Stephanie Efthymiou and Isabella Herman contributed equally to this study.
OI is a clinically and genetically heterogeneous connective tissue disorder hallmarked by increased susceptibility to bone fractures and is most commonly caused by monoallelic de novo pathogenic variants in COL1A1 (MIM: 120150) or COL1A2 (MIM: 120160). However, biallelic variants in genes involved in collagen type I biosynthesis have been frequently reported in consanguineous populations (Essawi et al., 2018;van Dijk et al., 2020;Van Dijk & Sillence, 2014). Currently, 20 different types of OI are identified in Online Mendelian Inheritance in Man (OMIM) (Amberger et al., 2015) with variable severity and phenotypic spectrum affecting primarily the skeletal system, although neurodevelopmental and other systemic complications have been observed in some autosomal recessive forms (e.g., MESD, MIM: 618644) (Moosa et al., 2019).
Here, we describe three affected children from two unrelated consanguineous families in order to expand the phenotype and further support the role of KDELR2 in AR OI. Informed consent, including consent to publish photographs, was obtained from the childrens' parents and institutional review board approval was obtained. All three children were clinically diagnosed with progressively deforming OI and neurodevelopmental delay. Three children had motor delay and two of three children had speech delay. The detailed clinical features of each patient are described in Table 1. Pedigrees, radiographs, and brain magnetic resonance images (MRIs) are shown in Figure 2. Common features observed in the affected patients include musculoskeletal abnormalities, including short stature and failure to thrive, Wormian bones, bowed limbs, chest deformity, hypotonia, joint hypermobility, and dysmorphic facies ( Figure 2). Family 1 consists of two affected children, a boy and a girl (P1, P2), born to consanguineous (first cousins) parents of Pakistani origin. Both patients have marked motor delay with inability to walk independently at 6 years and 2 years 8 months of age, respectively. The older child crawls as a means of ambulation and has never walked. He has had four fractures in his lifetime, the last at 4 years of age. The younger sister has not had any documented fracture to date at 2 years and 8 months of age.
She is not independently ambulatory but can take few steps with great support. In addition, she has speech delay with the first word spoken recently at 2 years of age. Common dysmorphic features in both siblings include epicanthus inversus, deep, sunken eyes, short neck, and thin, sparse hair. Brain MRI obtained from P1 at 6 years of age shows brachycephaly but is otherwise unremarkable (Figure 2(e)).
P3 was born to consanguineous first cousin Turkish parents with two prior miscarriages of unknown etiology. He was prenatally suspected to have OI due to ultrasounds showing abnormal bone structure. The patient has one unaffected sibling who does not carry the variant ( Figure 2). The patient's first fracture occurred at 21 days of age ( Figure 2(d)). Additional features observed include dentinogenesis imperfecta, blue sclera, scoliosis, and neurodevelopmental delay involving both motor and speech. Independent ambulation and speech emerged at 2 years of age; currently at age 4 years he is comparable to his neurotypical peers. Therefore, although he may have had early F I G U R E 1 KDELR2 loss of function (LoF) leads to inability of heat shock protein 47 (HSP47) to dissociate from procollagen. In wildtype cells, alpha collagen fibers assemble to form procollagen. Procollagen binds HSP47 and is transferred to the Golgi apparatus where KDELR2 binds HSP47 and leads to dissociation of HSP47 from procollagen. HSP47 is recycled back to the ER. Procollagen is further processed in the Golgi and secreted into the extracellular matrix (ECM) as tropocollagen. In mutant KDELR2 cells, KDELR2 is unable to bind HSP47.  Family-based exome sequencing (ES) with rare variant analysis was performed in both families followed by Sanger segregation for the identified variants as described before (Efthymiou et al., 2019;Manole et al., 2020). All three affected subjects were found to have  (Table 3), features which overlap with human biallelic pathogenic KDELR2 variants.
In conclusion, the data presented here support the role of KDELR2 in AR OI and expand the phenotypic spectrum of recessive KDELR2related AR OI first described by van Dijk et al. (2020) to include neurodevelopmental disorders such as motor and speech delay, as well as blue sclerae, dentinogenesis imperfecta, and hypotonia. However, motor delay and hypotonia are common features of OI and one reason they have not previously been reported may have been due to the small sample size of patients with this newly identified genetic etiology of OI. Additionally, it is unclear if the speech delay seen in early development is related to KDELR2, lack of exposure, or some other unidentified etiology. Noteworthy, the phenotypic spectrum of IMPCgenerated