Clinical, neuroimaging and molecular spectrum of TECPR2-associated hereditary sensory and autonomic neuropathy with intellectual disability

PURPOSE: Bi-allelic TECPR2 variants have been associated with a complex syndrome with features of both a neurodevelopmental and neurodegenerative disorder. Here, we aim to provide a comprehensive clinical description and variant interpretation framework. METHODS: Through an international collaboration, we identified 13 individuals from 12 families with bi-allelic TECPR2-variants. We systemically reviewed clinical and molecular data of this cohort and 11 cases previously reported. Phenotypes were standardized using Human Phenotype Ontology terms. RESULTS: A cross-sectional analysis revealed global developmental delay/intellectual disability, muscular hypotonia, ataxia, hyporeflexia, respiratory infections and central/nocturnal hypopnea as core manifestations. A review of brain MRI scans demonstrated a thin corpus callosum in 59%. We evaluated 14 distinct variants. Missense variants in TECPR2 are predominantly located in the N- and C-terminal regions containing {beta}-propeller repeats. Despite constituting nearly half of disease associated TECPR2 variants, classifying missense variants as (likely) pathogenic according to ACMG criteria remains difficult. We estimate a pathogenic variant carrier frequency of 1/1,221 in the general and 1/155 in the Jewish Ashkenazi population. CONCLUSION: Based on clinical, neuroimaging and genetic data, we provide recommendations for variant reporting, clinical assessment, and surveillance/treatment of individuals with TECPR2-associated disorder. This sets the stage for future prospective natural history studies.


INTRODUCTION
TECPR2 belongs to the tectonin β -propeller repeat-containing protein family and is implicated in the autophagy pathway. 1,2 Autophagy is critical to the development and function of the central nervous system. Loss-of-function variants in several genes of the autophagy pathway lead to both neurodevelopmental and neurodegenerative diseases. [3][4][5] In 2012, Oz-Levi et al. identified the homozygous TECPR2 variant c.3416del, p.(Leu1139Argfs*75) in five individuals from three Jewish Bukharian families and classified the syndrome as a novel subtype of hereditary spastic paraplegias (HSP) (SPG49; OMIM #615000). 6 To date, 11 individuals with bi-allelic TECPR2 variants have been reported in the literature. [6][7][8][9][10] All individuals showed muscular hypotonia and most had global developmental delay followed by intellectual disability. Only a subset of individuals displayed progressive spasticity as a characteristic HSP symptom. An autonomic and sensory neuropathy with respiratory, gastrointestinal and cardiovascular system involvement was present in a subset of individuals and central apnea was found to account for a large part of the morbidity. 7 indicated escape from nonsense mediated RNA-decay (NMD) but degradation of the truncated protein. 6 Functional data is largely missing for other described variants. This poses challenges for the interpretation of missense variants, for which normal expression of an altered protein is expected. All variants have been reported based on clinical overlap but have yet to be scored through the five-tier variant classification system recommended by the American College of Medical Genetics and Genomics (ACMG). 11 The lack of functional data and reliable variant classification have prevented an estimation of carrier frequencies and disease incidence, genotype-phenotype correlation analyses and the ability to make a genetic diagnosis in novel cases.
. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted October 20, 2020. . https://doi.org/10.1101/2020.10.10.20202622 doi: medRxiv preprint Through an international collaboration, we assembled a cohort of 13 individuals from 12 families of different ethnic backgrounds with known/novel disease-associated TECPR2variants. Based on a detailed review of the published cases and comparison with the herein described individuals, we provide a systematic clinical synopsis based on Human Phenotype Ontology (HPO). 12 We provide recommendations for surveillance and symptomatic treatment. An annotation and classification of all disease-associated variants according to the current ACMG recommendations is provided. 11 Using public databases, we estimate carrier frequencies and disease incidence. Based on this curated phenotype and genotype dataset, we propose a framework for reporting and validating TECPR2 variants.
. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

Review of Published Cases
A PubMed search identified five publications 6-10 describing 11 individuals from nine families diagnosed with TECPR2-associated disease (searched on 2020-09-10). Phenotypical features were extracted from published reports using the same questionnaire applied to novel cases.

Estimation of Carrier Frequencies from Public Databases
We retrieved all TECPR2 variants from gnomAD 25 and BRAVO (see Web Resources). These were annotated, scored, and filtered as described before to calculate carrier frequencies. 26

Analysis of Missense Variant Spectrum and Modelling of TECPR2 Protein Structure
Distribution of TECPR2 missense variants in the secondary protein structure was compared to missense variants reported as homozygous in public population databases and protein .
CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 20, 2020. . https://doi.org/10.1101/2020.10.10.20202622 doi: medRxiv preprint regions constrained for missense variation were analyzed as described before. 26 For analysis of the tertiary structure, we used the GalaxyWEB pipeline 27-29 to divide TECPR2 protein sequence into modeling units, predict their structure and refine the top model. Protein data bank (PDB) format structures (File S4) were then used for visualization with a pipeline using the Pymol software 30 and missense clustering analysis as described before. 26 For details, also see Supplementary notes.

RNA Expression Analysis for the c.2829del TECPR2 variant in P1
Messenger RNA from peripheral blood lymphocytes of P1 and both parents was used to generate cDNA. Monoallelic expression was analyzed with RT-PCR and Sanger sequencing ( Figure 2) and TECPR2 expression was analyzed using qPCR (see details in File S1).
. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 20, 2020. .

TECPR2 Variant Spectrum
Genetic analyses including conventional karyotyping, chromosomal microarray analysis and multi-gene panels (except for P13) were unremarkable in all novel cases. 14 distinct variants in TECPR2, including seven truncating and seven missense variants, were identified. Of these, three truncating and four missense variants have not been reported previously ( Figure   1A).

Founder Variants
The first reported founder variant 6  is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) 1 1 heterozygous with Ashkenazi founder variant). The variants are located in exons 5, 6, 7, 12 and 18. MAF in the heterozygous state was between 0 and 2/251,490 (gnomAD).

Expression analysis of the Stop Codon Containing Transcript in P1
Sanger sequencing of cDNA showed comparable detection of the normal allele and the allele with the c.2829del variant in both carrier parents of individual P1 (Figure 2A). Additionally, RT-PCR indicated normal expression in individual P1 who is homozygous for the c.2829del variant ( Figure 2B). Comparable expression of TECPR2 in individual P1, his parents and inhouse controls was confirmed by qPCR ( Figure S3).

Missense Variants
To date, only three missense variants have been reported (c. Analysis of spatial distribution in the linear protein structure indicated that missense variants identified in bi-allelic state in individuals with TECPR2-associated disease are predominantly located in the N-terminal (amino acid (AA) 1 to 357) and C-terminal (AA 802 to 1,411) protein regions. These two regions display a higher restrain for missense variation as indicated by higher computational scores and depletion of homozygous missense variants ( Figure 1A; Figure S1). . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 20, 2020. . https://doi.org/10.1101/2020.10.10.20202622 doi: medRxiv preprint This finding is further supported by the missense variant described in Spanish water dogs 31 , which is highly conserved (CADD PHRED score v1.6: 27.2) and located near to the c.4033G>C, p.(Ala1345Pro) variant (P3) in the C-terminal region (see Supplementary notes and Figure S1).
Our spatial proximity analysis using predicted 3D protein structures failed to identify clusters of missense variants (Table S3), but showed that all affect highly conserved residues in the repeats forming the N-terminal 7-bladed WD40 β -propeller or the two predicted C-terminal β propeller structures ( Figure 1B and 1C; Figure S2). While we choose the GalaxyTBM 28 model for visualization of the spatial missense distribution in Figure S1, the structural similarity of the model predicted de novo by the trRosetta algorithm 32 is remarkable ( Figure   S2, Table S2) . This conference of structure prediction algorithms add confidence to the derived models and will thus accelerate our understanding of missense variants in genetic disorders lacking experimentally derived protein structures.

Carrier Frequency for (Likely) Pathogenic TECPR2 Variants
Our results indicate that at least 1 in 1,221 individuals (0.082%) in gnomAD and 1 in 1,610 individuals (0.062%) in BRAVO is a carrier. In gnomAD we were able to estimate the carrier frequency for eight subpopulations, which ranged from 1 in 155 (0.650%; Jewish Ashkenazi) to 1 in 7,654 (0.013%; South Asian). Using these frequencies, the expected incidence is at least 1 in 5,961,640 newborns (based on gnomAD) to 1 in 10,366,419 newborns (based on BRAVO). Of the analyzed populations (which did not include the Jewish Bukharian population) the highest incidence is expected in the Jewish Ashkenazi population with 1 in 95,864 newborns.

Predicted Tertiary TECPR2 Protein Structure
The three different protein modelling algorithms that we have used, indicated similar results for the overall TECPR2 tertiary structure. The N-terminal domain (AA 1 to 357) containing seven WD-repeats is predicted to form a 7-bladed β -propeller fold (WD40 domain) with high . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted October 20, 2020.  Figure S2).

Clinical Spectrum
In our cohort of newly diagnosed cases, 10 of 13 individuals were male. Age at last follow up was between 16 months and 15 years with an average of 66.3 months. Consanguinity was reported in four out of the 11 families. Five families were of Jewish Ashkenazi descent, two families were of Jewish Bukharian. Except P1, all individuals were born at term without significant pre-or perinatal complications. Three individuals were small for gestational age.
Head circumference at birth was generally within normal limits. At last follow up only four individuals displayed short stature with a height below -2 SD (standard deviation) from agematched controls, however, all 11 individuals with data available were below average height.
Brachycephaly and microcephaly were observed in five individuals (three presented both).
Distinct facial features were seen in nine individuals though were not uniform. Shared characteristics included a short neck, synophrys and a triangular-shaped face. Skeletal abnormalities including significant lumbar kyphosis, a barrel-shaped chest or hyperextension of the neck were present in five cases.
The average age at diagnosis in our cohort ranged between 13 months and 15 years with an average of 47.5 months. 12 affected individuals showed global developmental delay and later intellectual disability (DD/ID) in the mild (n=1), moderate (n=5) and severe range (n=6).
P2 had only mildly delayed gross motor skills at last investigation; she is also the youngest individual in the cohort and developmental delay/intellectual disability may not be assessed . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted October 20, 2020. . https://doi.org/10.1101/2020.10.10.20202622 doi: medRxiv preprint accordingly due to her age. Four individuals with severe development delay were reported to show behavioral dysregulation with hyperactivity, restlessness, and overfriendliness. Two received a formal diagnosis of autism spectrum disorder. Eight children (age range: 16 months to 8 years) had not started walking at the time of last follow up and five individuals walked independently (average age: 45 months). P3 was diagnosed with dystonic/dyskinetic cerebral palsy and started walking around the age of 10 years. Speech development was delayed in all children and speech remained limited to a few words with two individuals remaining completely non-verbal.

Brain Imaging and EEG
Review of 12 brain MRI studies from our cohort (Figure 3 and Figure S4) and a review of reported cases in the literature defined a thin corpus callosum as a common feature (10/17, 59%). Additional findings in a subset of individuals included mild ventriculomegaly (often asymmetric colpocephaly), delayed myelination and diffuse cerebral atrophy. EEG was abnormal in three cases (3/11, 27%), but no specific pattern was reported.
. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

DISCUSSION
We here report a series of 13 individuals with bi-allelic TECPR2 variants from mostly nonconsanguineous families, and combine the detailed clinical, imaging, and molecular characterization of these individuals with the 11 cases previously reported. Since the variant c.2050C>G, p.(Leu684Val) was classified as likely benign according to ACMG criteria, one previously reported case 9 was excluded. The analysis of the remaining 23 individuals defines a core set of clinical and molecular features. These consist of global developmental delay and intellectual disability, axial and appendicular hypotonia, dysarthria and an abnormal gait, often described as an ataxic gait. Peripheral neuropathy was found in nearly all individuals in whom a detailed neurological assessment was available. Along with this, hyporeflexia was common and signs of autonomic dysfunction were prominent in the majority of cases. The latter included central hypoventilation, impaired temperature, and blood pressure regulation, repeat aspiration events and evidence of abnormal gastrointestinal motility. Whereas spasticity was recognized as a hallmark feature in the individuals initially reported 6 , the overall prevalence of spasticity was limited to a subset in our analysis (29%). We recognize that this is a potentially age-dependent manifestation, since increased tone was mainly reported in older individuals (P3 at age 15 years; Family B II-2 at age 20 years). P3 stands out because of the presence of dystonia, which was not present in previously published cases and possibly broadens the spectrum of neurological symptoms. Of note, epilepsy was nearly absent in our cohort, except in two individuals who experienced febrile seizures and two previously reported siblings with infrequent generalized tonic-clonic seizures. P13 stands out with therapy-resistant seizures, but due to consanguinity, other genetic conditions cannot be excluded. Future studies will be necessary to reassess epilepsy as associated feature.
Overall, the wide neurological manifestations in individuals with TECPR2-associated disease along the age spectrum, point to an involvement of multiple areas of the central nervous system (i.e. cortico-spinal tracts, cerebral cortex, brain stem, possibly basal ganglia) as well as the peripheral nervous system. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted October 20, 2020. . https://doi.org/10.1101/2020.10.10.20202622 doi: medRxiv preprint A large part of the morbidity and mortality associated with TECPR2 results from central hypoventilation requiring therapy with non-invasive positive pressure ventilation and occasionally active mechanical ventilatory support. Our findings are supported by a recently published, detailed analysis of the distinct breathing pattern from one affected individual. 10 Based on our clinical experience and the reported disease manifestations, we suggest a framework for routine surveillance as detailed in Table 2. Symptomatic treatment should be tailored to each individual case and aims at preserving function and preventing long-term morbidity and mortality. Early developmental support should be maximized to harness the developmental potential.
Overall, our cross-sectional analysis suggests that there is evidence of disease progression from a predominantly neurodevelopmental disorder with global developmental delay and hypotonia in early childhood to a progressive disease with corticospinal and corticobulbar dysfunction later on. We know from personal communications about the disease course of previously reported patients 6,7 , who all lost the ability to walk. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted October 20, 2020. . https://doi.org/10.1101/2020.10.10.20202622 doi: medRxiv preprint families with homozygous variants other than the founder variants, consanguinity of the parents was reported. This is exemplified for P1 where the run-of-homozygosity on chromosome 14 was not described in the CMA report, because it was below the 10 Mb filtering cutoff ( Figure 2C). Similar results were reported for P2 (File S1).
Our analysis did not show clustering or specific distribution pattern of the truncating variants.  Figure 1A; Figure S1 and S2), 2) deleterious effect predicted by in silico CADD score with cutoff >20 (PP3; Figure 1A), 3) the patient's phenotype matches the core features as well as TECPR2-specific symptoms of the HSAN-. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted October 20, 2020. . https://doi.org/10.1101/2020.10.10.20202622 doi: medRxiv preprint spectrum (Table 1) Table S1). Overall, based on these high carrier frequencies, both founder variants should be included in commercial carrier screening tests to inform genetic counseling and diagnostics in Jewish couples at increased risk for children with TECPR2-associated disease.
TECPR2 encodes a protein that is implicated in the early steps of the autophagy pathway where it interacts with the Atg8 family proteins, including LC3, to promote autophagic vesicle formation. 33 Fibroblasts from affected individuals showed a decreased number of autophagosomes and reduced delivery of LC3 and p62 for lysosomal degradation; this suggests an impairment of autophagic flux. 6 Providing insights into the mechanism of defective autophagy, a subsequent study showed that TECPR2 is involved in maintaining functional endoplasmic reticulum exit sites, which are implicated in the cargo from endoplasmic reticulum to Golgi and may serve as scaffolds for the formation of autophagosomes. 1 While the precise role of autophagy in TECPR2-associated disease remains to be established, there are several clinical features that are shared with other single gene disorders of this pathway. 3,4 This includes the involvement of multiple brain areas, clinical signs that point to a progressive involvement of the long CNS tracts, such as the cortico-. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted October 20, 2020. . https://doi.org/10.1101/2020.10.10.20202622 doi: medRxiv preprint 0 spinal tracts, as well as the imaging finding of a thinning of the corpus callosum. TECPR2associated disease, however, stands out for its prominent involvement of brain stem function, autonomic dysregulation, and peripheral neuropathy.
In summary, our cross-sectional analysis provides a depiction of clinical and molecular features across the age spectrum. Future prospective longitudinal studies are needed to better define the natural history and patterns of disease progression. Our present study provides a framework for assessing disease manifestations. Close follow up and surveillance for neurological and non-neurological manifestations is recommended.
. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted October 20, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted October 20, 2020. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted October 20, 2020. . https://doi.org/10.1101/2020.10.10.20202622 doi: medRxiv preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted October 20, 2020.   . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted October 20, 2020. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 20, 2020.  . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 20, 2020. . https://doi.org/10.1101/2020.10.10.20202622 doi: medRxiv preprint • Use sedatives with caution given reports of prolonged effects in this patient population . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 20, 2020. . https://doi.org/10.1101/2020.10.10.20202622 doi: medRxiv preprint