HACE1 deficiency leads to structural and functional neurodevelopmental defects

Objective We aim to characterize the causality and molecular and functional underpinnings of HACE1 deficiency in a mouse model of a recessive neurodevelopmental syndrome called spastic paraplegia and psychomotor retardation with or without seizures (SPPRS). Methods By exome sequencing, we identified 2 novel homozygous truncating mutations in HACE1 in 3 patients from 2 families, p.Q209* and p.R332*. Furthermore, we performed detailed molecular and phenotypic analyses of Hace1 knock-out (KO) mice and SPPRS patient fibroblasts. Results We show that Hace1 KO mice display many clinical features of SPPRS including enlarged ventricles, hypoplastic corpus callosum, as well as locomotion and learning deficiencies. Mechanistically, loss of HACE1 results in altered levels and activity of the small guanosine triphosphate (GTP)ase, RAC1. In addition, HACE1 deficiency results in reduction in synaptic puncta number and long-term potentiation in the hippocampus. Similarly, in SPPRS patient–derived fibroblasts, carrying a disruptive HACE1 mutation resembling loss of HACE1 in KO mice, we observed marked upregulation of the total and active, GTP-bound, form of RAC1, along with an induction of RAC1-regulated downstream pathways. Conclusions Our results provide a first animal model to dissect this complex human disease syndrome, establishing the first causal proof that a HACE1 deficiency results in decreased synapse number and structural and behavioral neuropathologic features that resemble SPPRS patients.

Homologous to the E6-AP carboxyl terminus domain and ankyrin repeat containing E3 ubiquitin-protein ligase 1 (HACE1), was identified to be downregulated in a number of tumors, as well as to play a role in inflammatory responses. [1][2][3][4] Analysis of Hace1 knock-out (KO) mice showed HACE1 is ubiquitously expressed, with relatively high expression in the brain. 2 A neurodevelopmental disorder named spastic paraplegia and psychomotor retardation with or without seizures (SPPRS) has recently been described in 15 patients of 6 unrelated families, associated with mutations throughout the HACE1 gene. [5][6][7] Mutations are predicted to be deleterious to HACE1 protein function, either disrupting folding or causing a frame shift and early termination of translation, resulting in no detectable protein product in patient-derived fibroblasts. 6,7 Clinical manifestations are variable and include early-onset developmental delays, severe intellectual disability, epilepsy, hypotonia, spasticity, ataxia, and difficulty with verbal communication. MRI findings include enlarged ventricles, hypoplastic corpus callosum, and altered white and gray brain matter ratios. There are several known targets for HACE1 ubiquitination, including Ras-related C3 botulinum toxin substrate 1 (RAC1), small GTPase, important for different aspects of brain development and neuronal function. 4,[8][9][10][11] In addition to cytoskeletal remodeling, RAC1 can regulate reactive oxygen species (ROS) levels as part of the nicotinamide adenine dinucleotide phosphate oxidase (NADPH) complex. 12 Correspondingly, Hace1 KO mice were reported to have elevated ROS levels. 13 While the role for hace1 was recently described in embryonic development in Xenopus laevis 14 and murine HACE1 has been implicated in Huntington disease, 13 nothing is known about the role of HACE1 during mammalian nervous system development or the molecular underpinnings of SPPRS. To elucidate the neurodevelopmental pathophysiology of SPPRS, we therefore performed detailed analysis of Hace1 KO mice and confirmed our findings in SPPRS patient-derived fibroblasts.
Standard protocol approvals, registrations, and patient consents Genetic and clinical data of patients 6-8 were published. 6 Our newly identified patients are labeled as patients [9][10][11]. Genetic data regarding patient 9 were produced as part of a clinical diagnostic service (GeneDx), and the research study including patients 10 and 11 has been approved by the ethics committee at The University of Göttingen. Biological materials of patients and healthy donors (HDs) and written informed consents were obtained in accordance with the Declaration of Helsinki. Fibroblasts from 2 female patients and 1 male patient only, 6, 7, and 10, respectively, are used in this study, as no consent was available for the others.

Mice
Animals were housed at the Institute of Molecular Biotechnology, Vienna, Austria, maintained under a 12-hour light/dark cycle, and provided with food and water ad libitum. Experiments described in this study were approved by the Bundesministerium fur Wissenschaft, Forschung und Wirtschaft (BMWFW-66.015/0004-WF/V/3b/2015) and performed according to EU-directive 2010/63/EU.

MRI
Anesthetized male C57Bl6/J mice 8-10 weeks of age and their Hace1 KO littermates were imaged in the Preclinical Imaging Facility at VBCF (pcPHENO, VBCF), member of the VBC, Austria, using a 15.2 T MRI (Bruker BioSpec, Glossary ANOVA = analysis of variance; HACE1 = HECT Domain And Ankyrin Repeat Containing E3 Ubiquitin Protein Ligase 1; HD = healthy donor; KO = knock out; LTP = long-term potentiation; NADPH = nicotinamide adenine dinucleotide phosphate oxidase; OFT = open-field test; PPI = pre-pulse inhibition; RAC1 = Ras-related C3 botulinum toxin substrate 1; ROS = reactive oxygen species; SPPRS = spastic paraplegia and psychomotor retardation with or without seizures; WT = wild type. Ettlingen, Germany). Values were normalized to brain size, averaged, and presented as percentage of total brain volume, and unpaired Student's t test was used to determine statistical significance.

Behavioral testing
Open-field test (OFT) and elevated plus maze were performed as described previously using an automated activity system (TSE-Systems, Germany). 15 Unpaired Student's t test was used to determine significance. Accelerating rotarod was performed on a 5-40 rpm accelerating apparatus (Ugo Basile), as described previously. 16 Significance was determined by 2-way analysis of variance (ANOVA) with Sidak's multiple comparisons. In the ladder rung walking task (pcPHENO, VBCF), mice were trained to walk across a ladder (length: 80 cm; spacing of rungs: 1 cm) toward an escape ladder in the TSE MotoRater system (tsesystems.com). Mice were recorded and manually scored using SimiMotion software 8.5.0.327. 17 Unpaired Student's t test was used to determine significance. T-maze was performed as described previously. 18 Significance of deviation of the number of successes in mutant vs wild-type (WT) mice was evaluated using a binomial test. For acoustic fear conditioning (pcPHENO, VBCF), mice were trained to associate the conditioned sound stimulus (CS = 85 dB, 10 kHz) with the unconditioned foot shock stimulus delivered by the floor grid (US = 1.5 mA), as described previously, using Coulbourn Habitest operant cages (Coulbourn Instruments, MA and FreezeFrame, Actimetrics, IL). 19 Unpaired Student's t test was used to determine significance. Mice were trained in the Morris water maze (pcPHENO, VBCF) as described previously. 20 Unpaired Student's t test was used to analyze the short-and long-term memory data and 2-way ANOVA with Sidak's multiple comparison test for the latency to reach the platform. Mice were tested for startle responses and prepulse inhibition (PPI) as described previously 21 using the SR-LAB Startle Response System (San Diego Instruments) chamber. Two-way ANOVA with Sidak's multiple comparison test was used to determine significance.
Hippocampal slices preparation and electrophysiologic recordings Electrophysiologic field recordings in the stratum radiatum of acute hippocampal slices were performed as described previously. 22 Two-population Student's t test were used to determine significance.

Synaptic number analysis
Hace1 KO; Thy1-green fluorescent protein (GFP) line M and hace1 WT; and Thy1-GFP line M 16 were analyzed using the LSM700 Axio Imager. Synapse number counts and measurement of neurite lengths were performed using ImageJ. 23 Statistical analysis was performed using Student's t test.
Active RAC1 pulldown Cellular levels of active RAC1 in fibroblasts were analyzed using the Active Rac1 Pull-Down and Detection Kit (Thermo Fisher Scientific) according to the manufacturer's instructions.

Migration assay
Primary dermal fibroblasts from patients 6, 7, and 10 and HDs were seeded at high density in 24-well plates for 24 hours and scratched using a pipette tip. The area of the scratch at 0 and 24 hours following the scratch was measured digitally in pixels, and the relative gap closure was calculated and normalized to wild-type cells. Unpaired Student's t test was used to determine significance.
ROS detection in the mouse brain and human primary dermal fibroblasts Twenty-micrometer coronal mouse brain sections were washed 2X in phosphate buffered saline and incubated in 10-mM dihydroethidium for 15 minutes at 37°C. Sections were washed twice in phosphate buffered saline and imaged with a Zeiss Axioplan2 microscope. Levels of ROS in fibroblasts were investigated using CellROX Deep Red staining (Thermo Fischer) followed by flow cytometric analysis according to the manufacturer's instructions. Statistical analysis was determined using unpaired Student's t test.

Results
Identification of novel HACE1 mutations in SPPRS patients Family 1 originating from Saudi Arabia had 1 affected female individual, patient 9, born to healthy, consanguineous parents (figure e-1A, links.lww.com/NXG/A148). The patient presented with similar clinical symptoms as previously reported cases of SPPRS, including severe intellectual disability, developmental delay, and inability to sit or speak by the age of 5 years. 6 Cranial MRI revealed microcephaly and brachycephaly, as well as hypoplastic corpus callosum and likely brainstem abnormality, small sella with ectopic neurohypophysis, and mild ventriculomegaly (figure e-1B). Additional clinical findings include mild facial dysmorphia, skeletal abnormalities, and ulnar deviation of the wrists and small feet (figure e-1C). In an unrelated consanguineous family 2 originating from Turkey, born to a healthy father and mother with Hashimoto disease are one female (patient 10) and one male (patient 11), 4 and 6 months of age at initial admission (Figure e-1D). During pregnancy, the mother was euthyroid and both children were carried to term. Patients had variable clinical symptoms in comparison with patient 9 and those previously reported, including hypotonia, small feet, enlarged head circumference, inverted and wide spaced nipples, facial dysmorphic features, and strabismus (figures e-1, E and F). Considering the pronounced hypotonia in all 3 patients, ataxia was difficult to determine. For detailed clinical symptoms of all patients, please refer to table.
Hace1-mutant mice exhibit brain morphologies similar to SPPRS patients To determine whether Hace1 KO mice develop any clinical symptoms of SPPRS patients, we first analyzed structural and cellular features of mutant mouse brains using 15.2 T MRI and/or histologic staining at different developmental stages. Histologic examination of mouse brains at embryonic day (E) 18.5, postnatal day (P) 0.5, and P14.5 revealed no evident defects cortical lamination, cerebellar foliation, hippocampal architecture, or cellular aberration and density in any developmental period examined (figure 1, B-D). However, enlarged ventricles and hypoplastic corpus callosum (figure 1, B and C) were noted in Hace1 KO animals as compared with WT littermates, as early as E18.5 and P0.5, respectively.
Western blot analyses of 8-week-old mouse brain lysates revealed HACE1 to be expressed throughout the adult mouse brain, which was absent in the brains of Hace1 KO mice (figure 2A). Total volume and morphology of the Hace1 KO adult mouse brain as compared with WT littermates measured by MRI and/or histologic examination revealed no whole-brain volumetric or cellularity differences as compared with WT littermates (figure 2, B and C). Review of histologic sections also revealed no abnormality in cortical laminar organization and cellular density in the adult brain ( figure 2D). Detailed analysis of different brain regions via MRI also showed no significant differences in the volumes of the cortex, colliculi, thalamus, and putamen, comparing adult WT and Hace1 KO brains, with a small decrease in the volume of the olfactory bulb (data not shown). The adult cerebellar volume and foliation as well as cortical laminar organization were comparable between Hace1 KOs and their control WT littermates (figure 2, E and F). Hippocampal volume was significantly decreased as measured by MRI ( figure 2G); however, there were no evident aberrations in cellularity and architecture on microscopic examination of hippocampi in adult animals ( figure 2H).
Of interest, volumetric MRI reconstructions revealed a significant reduction in whole-brain white matter (figure e-2, A-D, links.lww.com/NXG/A149) in Hace1 KO adult mice as compared with their WT littermates, substantiated by myelin basic protein staining and volumetric MRI analysis (figure e-2, B-D), similar to findings in reported SPPRS patients 5-7 and patient 9 described here (figure 1C and table). Also, noted was a reduction of the anterior commissure in Hace1 KO mice as compared with WT littermates (figure e-2C). In addition, in line with findings in SPPRS patients, as well as adult and  developing mouse brains, we observed enlarged ventricles in Hace1 KO mice as compared with control littermates ( figure  2C). These data show that Hace1 KO mice phenocopy key structural features associated with SPPRS patients.

HACE1-deficient mice exhibit locomotion and associative learning disabilities
To assess what, if any, behavioral defects can be detected in Hace1 KO mice, we subjected cohorts to an extensive battery of behavioral assays. The most notable clinical features of SPPRS are intellectual disability and ataxia 5-7 ; we therefore focused on hippocampal-and cerebellar-dependent tasks, principal brain regions regulating those behaviors in mice.
Hace1 KO mice generated in our group previously are generally healthy, viable, and fertile. 2 First, we found no significant differences in anxiety levels in Hace1 WT and KO littermates analyzed by the OFT and the elevated plus maze (figure e-3, A and B, links.lww.com/NXG/A150); however, we noted a reduction in the distance traveled during the testing (figure 3A). These data indicated that while our mice exhibited normal anxiety levels, there may be a locomotion defect. Therefore, we tested Hace1 KO mice for any locomotion and gait defects. WT and Hace1 KO mice were tested in the accelerating rotarod task, a standard assay to determine locomotion and gait defects. 16 While Hace1 KO mice and control littermates had similar latencies to fall off the rotating rod in the first trial, Hace1 KO mice exhibited shorter latencies to fall on all subsequent trials (figure 3B). We further analyzed the mice on a ladder rung walking task, which enables the observer to analyze coordinated front and hind paw placement on each individual rung of the ladder. 17 While Hace1 KO animals were capable of performing the task without falling off the apparatus and showed similar percentages of correct placement over all limbs analyzed, percent of major missteps made were significantly higher as compared with WT littermates. Both groups of animals had similar placement and correction scores, while WT animals had a higher percentage of partial placement on rungs (figure 3C). Thus, Hace1 KO mice exhibit locomotion and gait deficits.
The second notable clinical feature of SPPRS patients is intellectual disability (table). [5][6][7] To determine whether learning and memory processes are altered in Hace1 KO animals, we analyzed mutant mice in various hippocampal-dependent behavioral tasks. In the Morris water maze, KO mice were able to swim normally and there were no apparent differences in latencies between WT and Hace1 KO littermates to find the visible platform (not shown), in escape latencies during any of the training days (figure e-3C, links.lww.com/NXG/A150), nor in short-or long-term memory (figure e-3C, right panels). Likewise, there were no differences in T-maze between WT and Hace1 KO animals (figure e-3D). However, subjecting the animals to cued and contextual fear conditioning revealed significantly reduced freezing responses in the contextual aspect of the task in Hace1 KO mice compared with WT littermates ( figure 3D).
In addition, Hace1 KO mice exhibited significantly enhanced startle amplitude responses or reduced startle inhibition, in the acoustic startle task at presentation of 110-and 120-dB startle sounds as compared with their WT littermates ( figure  3E). These results also suggest Hace1 KO animals do not have any hearing deficiencies. PPI was reduced for all prepulse intensities presented (80, 85, 90, and 95 dB); however, this reduction was not significant when comparing Hace1 KO animals with WT controls (figure e-3E, links.lww.com/NXG/ A150). These data indicate that, besides locomotion defects, associative learning is altered in Hace1 KO mice, paralleling behavioral findings in human SPPRS patients.
Hace1 deficiency results in altered hippocampal synaptic transmission and reduced synapse numbers We and others have previously shown that altered synaptic transmission and plasticity at CA3-CA1 Schaffer collaterals can underlie deficits of contextual fear memory. 22,24 We tested whether synaptic transmission and plasticity of this pathway are affected in Hace1 KO mice. Recordings of field excitatory post synaptic potentials in response to CA3-CA1 synapse stimulation (schematic of the set-up in figure 4A) revealed an upregulation of excitatory synaptic transmission in hippocampal slices from Hace1 KO mice ( figure 4, A and B). Further analysis showed that the increase in synaptic transmission at the CA3-CA1 synaptic pathway was most likely due to a postsynaptic effect of the Hace1 deletion because neither the axonal fiber volley (figure 4C) nor the paired-pulse ratio, an indirect measure of the presynaptic release probability, 25 was affected (interstimulus interval 50 ms, paired-pulse ratio: WT 1.70 ± 0.066, n = 19, Hace1 KO 1.66 ± 0.036, n = 20, p = 0.64, 2-population t test, not shown). We then probed long-term potentiation (LTP), a cellular correlate of memory formation, 26 of this synaptic pathway and found that LTP is significantly reduced in acute hippocampal slices from Hace1 KO mice compared with WT littermates (figure 4, D and E). Of note, nerve conduction velocity was not affected in Hace1 KO recordings as compared with controls (data not shown), in line with the report that nerve conduction velocity is unaffected in SPPRS patients. 6 Because we observed hippocampal-dependent contextual fear learning deficiency and dysfunctions in electrophysiologic properties of pyramidal cells within the CA1 region of the hippocampus, we analyzed morphological properties of pyramidal neurons in this brain region. To this end, we crossed Hace1 KO mice with Thy1-GFP-M mice to drive neuronal GFP expression, which allowed us to visualize neuronal morphology, including spines along dendrites of CA1 pyramidal cells. 27 Quantification of selected dendritic segments indeed revealed a significant reduction in spine numbers in Hace1 KO neurons as compared with hippocampal neurons from WT littermates ( figure 4F). Whether such reduction in the synapse number is also present throughout the brain and might explain defective locomotion in Hace1 KO mice needs to be determined. These data indicate that, as a correlate to hippocampus-dependent impaired associate learning, HACE1 deficiency leads to a deficit of synaptic plasticity and a marked reduction of synapses in the hippocampus.

Elevated levels of active Rac1 in Hace1 KO mouse brains and SPPRS patient-derived fibroblasts
To determine molecular alterations downstream of HACE1 deficiency, we focused on Ras-related C3 botulinum toxin substrate 1, RAC1, a well-characterized substrate for HACE1mediated proteasomal degradation. 8,9 Indeed, Western blot analysis showed that RAC1 protein abundance was increased in Hace1 KO brains when compared with WT littermates ( figure 5A). In addition, Cyclin D1 abundance and ROS levels, as downstream read-outs for RAC1 activity, 12 were likewise increased in Hace1 KO brain regions as compared with controls ( figure 5, A and B).
The role of RAC1 has been extensively studied in the context of mouse brain development. 10,11 Of interest, perturbation of RAC1 levels, either overexpression or downregulation, results in decrease in dendritic spines in the mouse brain, similar to our Hace1 KO mice. 10,28 To determine whether RAC1 is also changed in HACE1-mutant SPPRS patients, we performed Western blotting analysis of SPPRS patient-derived fibroblasts, isolated from patients 6, 7, and 10 (table). Of note, in these patients, the mutations result in undetectable HACE1 protein expression (figure 5C), resembling our Hace1 KO mice. In all patients' fibroblasts, total abundance of RAC1 and Cyclin D1 was increased as compared with fibroblasts from HD controls (figure 5C). It is important that increased abundance of active GTP-bound RAC1 was detected in all 3 patient fibroblasts (figure 5D). As a functional consequence of increased RAC1 activity, ROS upregulation was also evident in SPPRS patient-derived fibroblasts ( figure 5E). In addition, in a scratch assay, patient fibroblasts migrated significantly faster, as compared with control fibroblasts (figure 5, F and G), confirming enhanced RAC1 activity. [29][30][31][32] These data show that loss of HACE1 expression results in increased levels of active RAC1 in both Hace1 KO mouse brains and HACE1mutant SPPRS patient-derived fibroblasts.

Discussion
Here, we report 2 novel mutations in HACE1, p.Q209* and p.R332*, discovered in 3 patients from 2 unrelated consanguineous families with a complex neurodevelopmental disorder. These new patients have variable clinical symptoms, overlapping with previous descriptions of SPPRS. [5][6][7] The genotype-phenotype correlations of the different mutations need to be addressed in future studies, as more SPPRS patients are diagnosed. We find that HACE1 is expressed throughout the adult mouse brain. Hace1 KO mice have defects in basic sensorimotor processing, deficiencies in specific learning and memory tasks, reduced LTP of hippocampal CA3-CA1 synapses, and fewer synaptic spines at CA1 pyramidal neurons. In addition, they exhibit locomotion defects as compared with littermate controls. Furthermore, we show a marked upregulation of RAC1 levels throughout the mutant mouse brain and elevated ROS levels. Verifying that our findings are relevant in humans, active RAC1 abundance, downstream signaling components, ROS production, and cellular migration in SPPRS patientderived fibroblasts are likewise dysregulated, all indicative of a hyperactive RAC1 pathway. Thus, HACE1 deficiency leads to neuroradiologic and behavioral manifestations in mice reminiscent of the clinical features seen in SPPRS patients.
A well-characterized target for HACE1 ubiquitination and subsequent proteasomal degradation is RAC1, a small member of the Rho family of GTPases, critical for neurogenesis, migration, axonal elongation, synaptogenesis, and activitydriven plasticity. 11 HACE1 is therefore poised to regulate many functions of RAC1, by modulating its active GTP-bound levels, including NADPH-dependent ROS production. 12,13 Mutations in several genes encoding guanine exchange factors and GTPase-activating proteins that regulate the on/off state of RAC1 have been identified to cause rare neuronal disorders. 33,34 Indeed, analysis of a protein-protein interaction subnetwork extracted from the human-integrated proteinprotein interaction reference database 35 reveals RAC1 to have 25 first-order interactors reported to be causative of intellectual disabilities. Neuronally expressed RAC1 first-order interactors are enriched in GTPase signal transduction by Gene Ontology analysis. Furthermore, RAC1 was also identified as a candidate gene for intellectual disability in a meta-analysis of several thousand trios, 36 and RAC1 missense mutations were recently reported to cause a neurodevelopmental disorder with variable symptoms in 7 individuals, some of which are overlapping with SPPRS patients. 37 Of interest, mutations in RAC1 caused remarkably different neurologic phenotypes that ranged from microcephaly to normal head circumference to macrocephaly, as well as variable intellectual disability (4/7), hypotonia (4/7), epilepsy (3/7), behavioral problems (3/7), and stereotypic movements (2/7). 37 The individuals who were determined to carry gain-of-function RAC1 mutations (3/7) had normal head circumference (1/3) or megalocephaly (2/3), in contrast to the Importantly, it has been previously established that dysregulation of RAC1 abundance in mouse models results in a decrease in dendritic spine number as well as reduction in hippocampal plasticity, as we have uncovered in our Hace1 KO mice. 10,28,38 We therefore speculate that RAC1 is a key factor underlining neuronal pathology in Hace1 KO mice and HACE1-deficient patients. Whether RAC1-dependent cytoskeletal modulation and/or ROS homeostasis is the key determinant feature of HACE1-dependent neurodevelopmental deficiencies remains to be elucidated. The present study adds HACE1 to the large RAC1 regulatory complex, which when perturbed has deleterious consequence to brain development and neuronal function. Continued