The Section on Molecular Dysmorphology studies the molecular, biochemical, and cellular processes that underlie rare genetic disorders due to impaired intracellular cholesterol homeostasis and perturbations of lysosomal function. These include malformation/cognitive impairment syndromes due to inborn errors of cholesterol synthesis and neurodegenerative disorders due to impaired intracellular cholesterol transport and lysosomal dysfunction. Smith-Lemli-Opitz syndrome (SLOS) and Niemann-Pick disease, type C1 (NPC1), respectively, are the prototypical examples of cholesterol homeostatic disorders whereas Juvenal Battens disease (CLN3) is an example of a lysosomal disease with unknown protein function. Other human malformation syndromes caused by inborn errors of cholesterol synthesis include Lathosterolosis, Desmosterolosis, X-linked dominant chondrodysplasia punctata type 2 (CDPX2), and Congenital Hemidysplasia with Ichthyosiform Erythroderma and Limb Defects (CHILD syndrome). Our section combines both basic and clinical research with the ultimate goal of developing and testing therapeutic interventions. Our basic research uses neuronal cell lines, mouse and zebrafish models of these disorders to understand the biochemical, molecular, cellular, and developmental processes that underlie the birth defects, learning and behavior deficits, neurodegeneration and a host of other clinical manifestations present in these various disorders. Our clinical research focuses on characterization and treatment of individuals with SLOS, NPC1 and CLN3. Our emphasis on both basic and clinical research allows us to integrate laboratory and clinical data in order to increase our understanding of the pathological mechanisms underlying SLOS and NPC1 to improve clinical care of these patients.
SLOS is an autosomal recessive, multiple-malformation syndrome characterized by dysmorphic facial features, intellectual disability, hypotonia, poor growth, and variable structural anomalies of the heart, lungs, brain, limbs, gastrointestinal tract, and genitalia. The SLOS phenotype is extremely variable. At the severe end of the phenotypic spectrum, infants often die as result of multiple major malformations. At the mild end of the phenotypic spectrum, SLOS combines minor physical malformations with behavioral and learning problems. The syndrome is attributable to an inborn error of cholesterol biosynthesis that blocks the conversion of 7-dehydrocholesterol (7-DHC) to cholesterol. Our laboratory initially cloned the human 3-beta-hydroxysterol delta7-reductase gene (DHCR7) and demonstrated pathogenic variants of DHCR7 in individuals with SLOS (Wassif et al., AM J Hum Genet. 1998; 63 (1);55-62). To date, together with others we have identified over 160 disease associated variants of DHCR7. We have further refined the incidence in SLOS to approximately 1:40,000 live births using a novel in silico approach mining large exome data sets (Cross et al. Clin Genet. 2015 Jun;87(6):570-5). To further our understanding of the mechanisms underlying the broad phenotypic spectrum in SLOS, we have developed iPS cells and discovered that 7-DHC leads to a defect in the Wnt/B-catenin pathway (Francis et al. Nat Med. 2016 Apr;22(4):388-96).
We also used gene targeting in murine embryonic stem cells to produce multiple SLOS mouse models, including a null deletion and a hypomorphic point mutation. Mouse pups homozygous for the null mutation (Dhcr7Δ3-5/Δ3-5) exhibit variable craniofacial anomalies, are growth-retarded, feed poorly, appear weak, and die during the first day of life because they fail to feed. Thus, we were not able to use them to study postnatal brain development, myelination, or behavior or to test therapeutic interventions. For this reason, we developed a mis-sense allele (Dhcr7T93M). The T93M mutation is the second most common mutation found in SLOS patients. Dhcr7T93M/T93M and Dhcr7T93M/Δ3-5 mice are viable and demonstrate SLOS with a gradient of biochemical severity (Dhcr7Δ3-5/Δ3-5>Dhcr7T93M/Δ3-5>Dhcr7T93M/T93M). We used Dhcr7T93M/Δ3-5 mice to test the efficacy of therapeutic interventions on tissue sterol profiles. As expected, dietary cholesterol therapy improved the sterol composition in peripheral tissues but not in the central nervous system. Treatment of mice with the statin simvastatin improved the biochemical defect in both peripheral and central nervous system tissue, suggesting that simvastatin therapy may be used to treat some of the behavioral and learning problems in children with SLOS. Most recently, we developed a zebrafish model for SLOS that will allow us to study the impact of aberrant cholesterol synthesis on behavior.
As part of our clinical studies on SLOS, we identified a novel oxysterol, 27-hydroxy-7-dehydrocholesterol (27-7DHC), derived from 7-DHC in SLOS patients. We therefore investigated whether 27-7DHC contributes to the pathology of SLOS and found a strong negative correlation between plasma 27-7DHC and cholesterol levels in these patients. In addition, previous work showed that low cholesterol levels impair hedgehog signaling. Therefore, we hypothesized that increased 27-7DHC levels would have detrimental effects during development in response to suppression of cholesterol levels. To test our hypothesis, we produced SLOS mice (Dhcr7Δ3-5/Δ3-5) expressing a CYP27 (sterol 27-hydroxylase) transgene. CYP27Tg mice display increased CYP27 expression and elevated 27-hydroxycholesterol levels but normal cholesterol levels. While Dhcr7Δ3-5/Δ3-5 mice are growth-retarded, exhibit a low incidence of cleft palate (9%), and die during the first day of life, Dhcr7Δ3-5/Δ3-5:CYP27Tg embryos are stillborn and have multiple malformations, including growth retardation, micrognathia, cleft palate (77%), lingual and dental hypoplasia, ankyloglossia, umbilical hernia, cardiac defects, cloacae, curled tails, and limb defects. We observed autopod defects (polydactyly, syndactyly, and oligodactyly) in 77% of the mice. Consistent with our hypothesis, sterol levels were halved in the liver and 20-fold lower in the brain tissue of Dhcr7Δ3-5/Δ3-5:CYP27Tg than in Dhcr7Δ3-5/Δ3-5 embryos. The fact that 27-7DHC plays a role in SLOS may explain some of the phenotypic variability and may lead to development of a therapeutic intervention. The project is a good example of the benefits of integrating clinical and basic science to both understand the pathology of SLOS and develop potential therapeutic interventions.
We continue conducting a longitudinal Natural History trial with a new focus on older SLOS patients as treatment and improvements in medical care have generated a population of patients that has been previously difficult to ascertain. Given that SLOS patients have a cholesterol deficiency, they may be treated with dietary cholesterol supplementation. We have completed the first double blinded cross over control study for the use of simvastatin as a therapeutic intervention in SLOS in a group of 18 patients. We found significant improvements in serum 7-DHC levels, a trend towards improvement in the central nervous system, and behavioral improvements.
One reason for studying rare genetic disorders is to gain insight into more common disorders. Most patients with SLOS exhibit autistic characteristics. We are currently collaborating with other NIH and extramural groups to evaluate this further.
Lathosterolosis, Desmosterolosis, and HEM dysplasia
Lathosterol 5-desaturase catalyzes the conversion of lathosterol to 7-dehydrocholesterol, representing the enzymatic step immediately preceding the defect in SLOS. Thus, to gain a deeper understanding of the roles of reduced cholesterol versus elevated 7-dehydrocholesterol in SLOS, we disrupted the mouse lathosterol 5-desaturase gene (Sc5d) by using targeted homologous recombination in embryonic stem cells. Sc5d−/− pups are stillborn, present with micrognathia and cleft palate, and exhibit limb-patterning defects. Many of the malformations in the mutant mice resemble malformations in SLOS and are consistent with impaired hedgehog signaling during development. Biochemically, the mice exhibit markedly elevated lathosterol levels and reduced cholesterol levels in serum and tissue.
A goal of producing a lathosterolosis mouse model was to gain phenotypic insight for the purpose of identifying a corresponding human malformation syndrome. We identified a human infant patient with lathosterolosis, a malformation syndrome not previously described in humans. Biochemically, fibroblasts from the patient show reduced cholesterol and elevated lathosterol levels. Mutation analysis showed that the patient is homozygous for a single A→C nucleotide change at position 137 in SC5D, resulting in a mutant enzyme in which the amino acid serine is substituted for tyrosine at position 46. Both parents are heterozygous for the mutation. The infant’s phenotype resembled severe SLOS. Malformations found in both the human patient and the mouse model include growth failure, abnormal nasal structure, abnormal palate, micrognathia, and postaxial polydactyly. A unique feature of lathosterolosis is the clinical finding of mucolipidosis in the affected infant, which is not reported in SLOS and may help distinguish SLOS clinically from lathosterolosis. Lathosterolosis, which is a lysosomal storage disorder, may be replicated in embryonic fibroblasts from the Sc5d–mutant mouse model. To distinguish pathological changes attributable to reduced cholesterol in lathosterolosis from those that are a consequence of elevated 7-DHC in SLOS, we are comparing proteomic changes in the Sc5d–mutant mouse model with those in the SLOS mouse model. We recently developed induced pluripotent stem cells from lathosterolosis fibroblasts. Their characterization is in progress.
Desmosterolosis is another inborn error of cholesterol synthesis that resembles SLOS. It results from a mutation in the 3β-hydroxysterol Δ24-reductase gene (DHCR24). DHCR24 catalyzes the reduction of desmosterol to cholesterol. We disrupted the mouse Dhcr24 gene by using targeted homologous recombination in embryonic stem cells. Surprisingly, although most Dhcr24 mutant mice die at birth, the pups are phenotypically normal.
Others have shown that mutations of the lamin B receptor (LBR) cause HEM (hydrops, ectopic calcification, moth-eaten skeletal) dysplasia in humans and ichthyosis in mice. LBR has both lamin B–binding and sterol Δ14-reductase domains. Although only a minor sterol abnormality has been reported, it was proposed that LBR is the primary sterol Δ14-reductase and that impaired sterol Δ14-reduction underlies HEM dysplasia. However, DHCR14 also encodes a sterol Δ14-reductase. To test the hypothesis that LBR and DHCR14 are redundant sterol Δ14-reductases, we obtained ichthyosis mice (Lbr Sc5d−/−) and disrupted Dhcr14. Dhcr14 Sc5d−/− mice are phenotypically normal. We found no sterol abnormalities in either Lbr Sc5d−/− or Dhcr14 Sc5d−/− tissues at 1 or 21 days of age. We then bred the mice to obtain compound mutant mice. Lbr Sc5d−/−:Dhcr14 Sc5d−/− and Lbr Sc5d−/−:Dhcr14+/− died in utero. Lbr+/−:Dhcr14 Sc5d−/− mice appeared normal at birth but, by 10 days of age, were growth-retarded and neurologically abnormal (with ataxia and tremors) and, consistent with a demyelinating process, evidenced vacuolation and swelling of myelin sheaths in the spinal cord upon pathology evaluation. We observed neither vacuolation nor swelling of myelin sheaths in either Lbr Sc5d−/− or Dhcr14 Sc5d−/− mice. In contrast to Lbr Sc5d−/− mice, Lbr,+/−:Dhcr14 Sc5d−/− mice had normal skin and did not display the Pelger-Huët anomaly. Peripheral tissue sterols were normal in all three mutant mice, although we found significantly elevated levels (50% of total sterols) of cholesta-8,14-dien-3β-ol and cholesta-8,14,24-trien-3β-ol in brain tissue from 10-day-old Lbr+/−:Dhcr14 Sc5d−/− mice. In contrast, we observed relatively small transient elevations in Δ14-sterol levels in Lbr Sc5d−/− and Dhcr14Δ4-7/Δ4-7 brain tissue. Our data support the notion that HEM dysplasia and ichthyosis result from impaired lamin B receptor function rather than from impaired sterol Δ14-reduction. Impaired sterol Δ14-reduction gives rise to a novel murine phenotype for which a corresponding human disorder has yet to be identified.
Niemann-Pick type C
Niemann-Pick disease, type C1 (NPC) is a neurodegenerative disorder characterized by progressive cerebellar ataxia and dementia. Mutation in either the NPC1 or NPC2 gene result in perturbations of intracellular lipid and cholesterol transport and ultimate storage in the late endosome/lysosome. This aberrant storage results in a cascade of cellular and clinical deficits resulting in loss of neurons and ultimately death. A natural history/observational trial was initiated in 2006 with the goal of identifying biomarkers and clinical features that could be used as outcome measures in therapeutic trials. This work has been supported by NIH Clinical Center Bench-to-Bedside awards the Ara Parseghian Medical Research Foundation and Dana’s Angels Research Trust. We have enrolled more than 120 NPC1 patients in this longitudinal trial. The goals of the trial at initiation were to identify (1) a blood-based diagnostic/screening test, (2) biomarkers that can be used as tools to facilitate development and implementation of therapeutic trials, and (3) clinical symptoms/signs that may be used as efficacy outcome measures in a therapeutic trial. We have successfully in collaboration with Daniel Ory (University of Washington) found elevated levels of non-enzymatically produced oxysterol, cholestane-3β,5α,6β-triol, in NPC1 patients. This biomarker has been successfully developed into a CLIA approved blood-based assay for NPC and is now commercially available. A bile acid derivative of this oxysterol is being tested for its utility in a newborn screen. We have additionally identified several cerebral spinal fluid biomarkers such as calbindin D and fatty acid binding protein 3 that will help in the evaluation of experimental therapeutic interventions.
In addition to our Natural History study, in collaboration with the Therapeutics of Rare and Neglected Disease Program of NCATS, we have completed a Phase 1/2a trial of lumbar intrathecal 2-hydroxypropyl-β-cyclodextrin (HPβCD) therapy in NPC (Ory et al. Lancet 2017 Oct 14;390(10104):1758-1768., and Berry-Kravis et al. Pediatr Neurol. 2018 Mar;80:24-34.) and a multicenter, multinational Phase 2b/3 trial. In collaboration with Daniel Ory and Frederick Maxfield, our group was awarded an NIH U01 grant to test the safety and potential efficacy of an HDAC inhibitor, vorinostat, in adult NPC1 patients. The collaboration also includes scientists from Notre Dame and has been supported by the Ara Parseghian Medical Research Foundation. It is also crucial we continue to evaluate existing therapies. Miglustat, a substrate reduction therapy while not FDA approved to treat NPC in the United States has been approved and widely used as a therapeutic intervention in Europe. Approximately half of our patients are currently on miglustat which we have recently demonstrated improves swallowing (Solomon BI, et al. JAMA Neurol. 2020 Dec 1;77(12):1564-1568.)
As therapeutic trials become available it is critical to ascertain the number of individuals affected by this disease and identify them early when clinical intervention maybe most beneficial through use of biomarkers for identification and assessing the efficacy of therapeutic interventions. Using the same approach, we employed in SLOS we verified an incidence of NPC1 of approximately 1:89,000 live births, however we identified two possible higher frequency variants that may increase the incidence of NPC1 to as high as approximately 1:39,000 live births. We are actively investigating and validating biomarkers using a large assortment of platforms.
We have initiated gene therapy as a treatment of NPC1 in collaboration with Charles Venditti and William Pavan both in NHGRI. Using a NPC1 construct in an adeno-associated vector system we have been able to extend the life span of the NPC null mouse. This work is ongoing.
We continue to collaborate with laboratories throughout the world looking at aspects of the disease such as Tau accumulation in the NPC1 mouse brains and CSF of our patient cohort, exploring the dysregulation of the P450 system in both mice and men, continuing to further refine the composition of the stored lipids in NPC1 lysosomes, attempting to determine the functional manner in which HPβCD interacts within or on a cell to release the stored material from the lysosome, and continue to explore multiple therapeutic interventions. To further complement the clinical work, we have begun to develop zebrafish models of NPC1 and NPC2. These models will enhance and increase the rate we can investigate therapeutic interventions as well as provide additional avenues to investigate the pathophysiology of these disease.
Juvenile Batten Disease
Ceroid lipofuscinosis, neuronal, 3 disease (CLN3 disease, Juvenile Batten disease, OMIM#204200) is a rare, fatal, neurodegenerative lysosomal disease caused by mutation of CLN3. JNCL is one of a group of 13 disorders categorized as neuronal ceroid lipofuscinoses. In aggregate these are considered the most common neurodegenerative disorders in children with incidence estimates ranging from 1/12,500 to 1/100,000 in European and USA populations. The incidence of CLN3 disease varies with the study populations but the worldwide incidence is approximately 1 in 100,000 live births. Neurological symptoms typically manifest between 4 and 7 years of age. The initial clinical manifestation of CLN3 is rapid, progressive vision loss followed by cognitive impairment and behavioral disturbances. Progressive neurological impairment includes motor dysfunction and seizures followed by death as young adults. There are no approved therapies for CLN3 disease.
CLN3 maps to chromosome 16p12.1 and encodes a 438 amino acid integral lysosomal membrane protein. Mutations of CLN3 underlie this disorder. The most common mutation of CLN3 is a 966 bp deletion (“1 kb deletion”) of exons 7 and 8. Biochemically, CLN3 is characterized by lysosomal storage, the accumulated material contains significant amounts of subunit c of F1F0ATP synthase complex (SCMAS), but is also rich in lipids such as lysobisphosphatidic acid, consistent with an endosomal/lysosomal origin. Although multiple cellular functions of CLN3 have been proposed, the function of CLN3 remains unknown. Determining the function of CLN3 and developing therapies has been hampered by the lack of model systems that fail to adequately recapitulate the human disorder. Understanding the normal function of CLN3 and its role in disease pathology is critical to the identification and development of therapeutic interventions.
To facilitate study of CLN3 function and provide a tool for high-throughput chemical and genetic screens we are using a recently developed human induced pluripotent stem cell (iPSC) line referred to as i3neurons (inducible × integrated × isogenic). The i3neuron platform is a scalable iPSC-derived neuron technology that allows for the reliable and reproducible generation of highly pure human neurons in vitro and is therefore amenable to high-throughput chemical and genetic screens.
In addition to our Natural History study, we have initiated a trial studying the benefits of using an optic-to-audio assistive device. These studies are laying the groundwork for biomarker discovery with similar goals as our NPC1 studies and for improvements to patients daily lives.