Abstract
Two siblings with an early onset of a neurodegenerative disease were presented with muscular hypotonia, secondary microcephaly, and severe developmental delay. Seizures were refractory to treatment but could be controlled with a ketogenic diet. Over the course of 5 years, whole exome sequencing (WES) was performed twice in both children. The first time the diagnosis was missed. The next one revealed compound heterozygous mutations in the gene coding for the tubulin folding cofactor D. Technical improvements in WES mandated a new investigation after a few years in children where the diagnosis has not been found.
Introduction
Microtubules play a crucial role for the cytoskeleton and are essential to develop synaptic connections, as well as to build axons and dendrites. They consist of α- and β-tubulins that form heterodimers and subsequently protofilaments. Several protofilaments build a hollow tubular structure. Polymerization and depolymerization of the microtubules are essential in proliferation, communication, development, and motility, which are organized by microtubule-organizing centers (MTOCs). To settle the polarity of the microtubules, a γ-tubulin ring complex is crucial because it functions as a microtubule nucleator at the MTOC and binds the β-subunit to expose the α-subunit (Zheng et al 1 ).
The correct building of microtubule blocks requires additional proteins and cofactors. Native α- and β-tubulins initially interact with prefoldin, which is responsible for bringing them together with their corresponding chaperonin (chaperonin containing TCP-1 [CCT]—cytosolic chaperonin) (Vainberg et al 2 ). After adenosine triphosphate hydrolysis, α- and β-tubulins interact with specialized chaperons, the tubulin folding cofactors, which assist in the correct folding of the α/β-tubulin heterodimers. Tubulin folding cofactor A (TBCA) is part of the β-folding pathway and tubulin folding cofactor B (TBCB) is part of the α-folding pathway. Tubulin folding cofactor C (TBCC), tubulin folding cofactor D (TBCD), and tubulin folding cofactor E (TBCE) are involved in both pathways. TBCB and TBCA need to bind to native α- and β-tubulins before the following cofactors can build a complex. Cofactors D, E, and C build the super complex with α- and β-tubulin to hydrolyze guanosine triphosphate (GTP). This results in the release of the correctly folded native α- and β-tubulin (see Figure 1). α- and β-tubulin monomers from depolymerized microtubules may also be recycled into this pathway by direct binding to TBCE and TBCD, respectively (see dotted dash line arrows in Figure 1) (Lewis et al 3 ). ARL2, an adenosine diphosphate ribosylation factor-like protein, is involved in the regulation of the folding pathway by interacting with TBCD and β-tubulin. The TBCD·ARL2·β-tubulin trimer is a functional complex whose role is crucial in microtubule assembly (Francis et al 4 ).
(CCT) This picture shows the tubulin folding pathway and the involvement of several cofactors. The microtubules depolymerization depends on GTP hydrolysis and the correct interaction with the cofactors. TBCA and TBCB are described exclusively in 1 pathway, whereas TBCC, TBCD, and TBCE build the super complex to hydrolyze GTP and facilitate the release of the native tubulins. TBCD is crucial to release the native β-tubulin (modified from Bhamidipati et al 5 ).
Several neurodegenerative diseases are described in the literature that have their cause in mutations of proteins, which are involved in microtubule formation or posttranslational modifications 6 (MIM: #617193). Mutations in TBCD can result in a defective function of the cofactor and deficiencies in the structure, folding, or stability of the microtubules 7 (Flex et al 8 ) and lead to neuronal degenerative disorders (Miyake et al 9 ).
Whole exome sequencing (WES) is a very powerful method to diagnose ultra-rare genetic disorders such as TBCD deficiency. Often, it is only done once and not repeated even if no diagnosis could be found. However, continuous improvement of the method suggests that it can be useful to repeat the analysis.
Materials and Methods
Array-comparative genomic hybridization (CGH) was performed in February 2010 using an Agilent 400k Array. Single nucleotide polymorphism (SNP)-chip analyses for homozygosity and identical by descent mapping were done essentially as described in Tegtmeyer et al 10 using Illumina 660 W beadchips. The logarithm of the odds scores over the chromosomes were calculated using Merlin-1.1.2 software. 11 Two exome analyses were done by CEGAT (Tübingen). For the first exomes in 2013, enrichment of exonic sequences was achieved with the Agilent Exome v.5 kit. For sequencing, an Illumina HiSeq2500 was used. For each patient, data from two runs were pooled. Reads were mapped to the hg19 reference sequence using the bwa-tool 0.7.2-r351. Variants were called with samtools 0.1.18 + bcftools 0.1.17 and VarScan 2.3.5. Annotation was based on Ensembl.74 + RefSeq.UCSC.20131210 + CCDSr15, dbSNP 138, dbNSFP 2.1, exome variant server , and human gene mutation database (HGMD) version 2013.04. Regarding the new exomes in 2018, enrichment was by Agilent Exome v.7, adapter trimming by Skewer 0.2.2 software, mapping to hg19 reference sequence using the bwa-tool 0.7.2-r351, variant calling by CeGaT StrataCall r1254, annotation was based on Ensembl.75 + RefSeq.UCSC.20180712 + CCDSr15, dbSNP 151, dbNSFP 3.4c, GnomAD 2.0.2, HGMD 2018.02, and indel realignment by assembly based realigner (PMID: 24907369). 11
To investigate the effect of the discovered mutations in the TBCD gene (RefSeq GenBank: NM_005993.5), complementary DNA (cDNA) was synthesized using RNA obtained from peripheral blood cells by standard procedures. The region of interest was amplified using primers in exons 20 and 28 (5′- GAGGCTGCTGTCCATGACACTGAG, 5′-GGGGAAGTGTGGGTAACTGCTCT, product length 606 bp). Primers were 5′-tailed with M13 forward and reverse adapters for simplified sequencing M13F: TGTAAAACGACGGCCAGT and M13R: CAGGAAACAGCTATGACC. The product was sequenced by standard Sanger sequencing, using BigDye-terminator technology and primers M13F and M13R. Primers were searched using Primer blast software 12 against sequences of the respective exons with neighboring intron regions. These sequences were extracted from Ensembl, ENSG00000141556.
Results
Two siblings with a neurodegenerative disease are described with epileptic seizures, secondary microcephaly, and a severe developmental delay. Both siblings carry heterozygous, disease-causing mutations in the TBCD gene.
Patients
The two siblings descent from healthy, unrelated parents who did not report any inherited disease in their family. Informed consent was obtained from the parents.
Patient A (female) was born 3 weeks prior to the estimated date of birth with a body weight of 3140 g (36th percentile), length of 51 cm (48th percentile), 33.5 cm head circumference (24th percentile), Apgar 09/10/10, and umbilical cord pH (A) 7.3. She was presented at our hospital at the age of 8 months due to seizures and developmental delay. During the seizures, she grimaced and gazed at the top right. She showed hypersalivation and a fencing posture. The seizures were initially fever-associated but later occurred unprovoked. In the first clinical examination, she showed few spontaneous movements and could not sit without support. She presented with a progressive hypotonia of the lower limb that had been noticed before in a routine screening at the age of 6 months. She had a secondary microcephaly as well as a marginal splenomegaly with 7.5 cm (normal: <7.2 cm). No cardiac, abdominal, or pulmonic abnormalities were observed at this point of the investigation. She developed talipes equinus on both sides and hip luxations. Her length and weight were within the 25th and 50th percentile but her head circumference was below the third percentile. The newborn screening for metabolic disorders and the blood amino acids were normal. During the first 9 months, methylmalonic acid in the organic acids was several times significantly elevated and declined after the administration of vitamin B12.
The first electroencephalogram (EEG) was pathological with beta activity overlay and epileptiform discharges on the right hemisphere. Physiological stages of sleep were not verifiable. The following EEGs showed sharp-wave complexes on the left and right hemispheres. Tachycardia and low oxygen saturation were part of the seizures and generally higher amplitude on the right hemisphere was noticed. The patient slept a lot and regularly showed small seizures after waking up during the day. The seizures were both myoclonic and tonic and occurred with screaming or laughing.
Several anticonvulsive drugs were used in mono and combination therapy (eg sulthiame, oxcarbazepine, topiramate, ethosuximide, and levetiracetam) led to fewer seizures, however, episodes with atypical absences remained. Finally, the patient was free from seizures after a ketogenic diet was introduced (see Figure 2).
Seizure frequency under anticonvulsive therapy and ketogenic diet in patients A and B.
Brain magnetic resonance imaging (MRI) revealed reduced supratentorial white matter volume and a hypoplasia of the corpus callosum. The neurocranium and myelination were age-appropriate at 8 months; however, a lack of developmental progression of myelinization in the following MRI was observed. Progressive brain atrophy with consecutive hydrocephalus ex vacuo was seen in the subsequent MRIs. Brain atrophy affected gray and white matter equally including the basal ganglia. The head of the caudate nucleus was almost undetectable in the MRI (see Figure 3).
Patient A (1-3): Magnetic resonance imaging at the age of 12 months with progressive loss of brain volume. Extended liquor ventricles, hypomyelination and progressive depletion of the corpus callosum, and an almost nondefinable head of caudate nucleus (arrow) were observed. The photograph (4) was taken at the age of 7 years. Patient B (5-7), Magnetic resonance imaging at the age of 11 months showing a distinct global brain atrophy, hypomyelination and a thin corpus callosum. Especially the hippocampal regions and the caudate nuclei are thinned out (arrow). The photograph (8) was taken at the age of 8 years.
Motor nerve conduction velocity was significantly reduced since the patient's first presentation at the age of 8 months (nervus tibialis posterior and nervus ulnaris). No signs of inflammation were observed in the muscle biopsy, but a pattern of neurogenic damage with fiber atrophy and hypertrophic groups of fibers was found. Sural nerve biopsy was normal. Her patellar, Achilles tendon, biceps, and triceps tendon reflexes were present until the age of 2.5 years. At this point exclusively the pupillary light reflex was consistently noticeable. She expressed almost no control over her head movements and developed a heavy tremor. She showed little motoric development but was able to express emotions such as joy and pain.
At 19 months of age, a premature thelarche was reported. Her breast development was classified according to the Tanner System as PH 1 B3. Furthermore, she had a growth spurt without a puberty-typical hormone profile or bone age, which made a precocious puberty seem unlikely. Due to muscular hypotension, she suffered from dysphagia and recurrent pulmonary infections. Furthermore, she developed an elevated diaphragm and atelectasis, which was treated symptomatically. Both patients developed intermittent neutropenia in the first years of life.
She had not reached the development milestones appropriate for her age from the age of 7 months onward, but she even lost previously acquired capabilities. At the age of 10 months, she could control her head occasionally, but she could not roll from front to back on her own, neither could she sit freely. Her grasp was uncoordinated, but she could laugh, focus, babble unspecifically, and show emotions.
She died at the age of 7 years of an airway infection.
Her younger brother, patient B, was born 2 weeks prior to the estimated date of birth with a body weight of 3420 g (52nd percentile), length of 50 cm (26th percentile), head circumference of 35 cm (52nd percentile), Apgar 09/10/10, and umbilical cord pH (A) 7.3. He had developed completely appropriate to his age in the first 3 months after birth. Then he started to move less and showed extension in the legs while lifting him. At the age of 6 months, he presented with a hypotonia of the lower limb after kicking and moving actively in the first months after birth. He showed progressive hypotonia, neurodevelopmental delay, and fever associated epileptic seizures. In the first examination, the muscular reflexes were still present. By 9 months he stopped turning on the sides. He was diagnosed with a massive psychomotor retardation.
The EEG showed an asymmetry between the right and the left hemisphere. Sharp waves and beta activity overlay were observed. A treatment with levetiracetam did not decrease the seizure frequency, and he stayed very tired during the day. By 12 months he presented with fever while seizures appeared every 3 min. The anticonvulsive therapy was optimized with phenobarbital, which could decrease the seizure frequency to once every 20 to 60 min. Potassium bromide was added on suspicion of migrating partial epilepsy but needed to be stopped because of side effects. The seizure frequency remained high during occasionally occurring fever episodes. Eventually, a ketogenic diet finally decreased the seizure frequency with a fat–carbohydrate ratio of 3:1. The ketone bodies in the urine exceeded 80 mg/dL (see Figure 2).
During the day, he still had several myoclonic twitches lifting his arms and opening his eyes wide. Occasionally he expressed a fencing posture. The pattern of myoclonia presented identically compared with the sister’s myoclonia. No physiological sleep structures appeared in the EEG, and he maintained a high susceptibility to seizures. The MRI results of both siblings showed similarities (see Figure 3). He developed dysphagia and from the age of 2.5 years, he received a majority of his food via a percutaneous endoscopic gastrostomy tube. He did not develop any ability to speak, had convergent strabism, and at 9 years old he developed nystagmus. The parents reported that the hand-mouth-contact had stopped by 5 years, and by 7 years he showed significantly less movement and anticipation. He is now 10 years old.
Retrospectively, the mother reports several differences among the pregnancies with the 2 affected children and the following healthy sibling. The affected pregnancies shared a cervix insufficiency, which required a cervical cerclage; increased heart rates of the fetuses at around 160/170 bpm; and significantly less movement and activity. An ultrasound in the 20th week already showed borderline abnormalities in the brain, and both affected children were born 3 weeks early due to premature contractions.
Genetics
The chromosomal analysis showed a standard karyotype and array CGH was inconspicuous. Since the parents came from a limited population of German immigrants to Russia, we also checked for unknown consanguinity by SNP-chip-analysis. In total 13.7 Mb of the genome were shared homozygous by the siblings, but WES showed no relevant mutations in those regions. About one-third of the genome was shared identical by descent by the siblings, but using the results of the first exome analysis no compound heterozygous mutations led to a candidate gene. The average coverage of that exome was 144× with 96.2% at least 30× for patient A and 308× with 98.7% at least 30× for patient B. The second exome in 2018 revealed 2 mutations in TBCD in both siblings (see Table 1). Retrospective inspection of the 2013 data showed that the c.230A > G was contained in the variant lists of both siblings, but the gene did not appear as relevant since the second mutation was not covered sufficiently (see Table 1, numbers taken from mapping-file), was therefore filtered off and was not listed in the respective variant list of patient A.
The Coverage of the Correspondent Regions in Whole Exome Sequencing (Patient A Female, Patient B Male); Given are the Counts Normal/Mutant Allele.
Furthermore, no disease-causing mutation had been identified and published until 2016. The WES in 2018 (average coverage 149×; 96.7% at least 30× for patient B) finally led to our diagnosis and located the compound heterozygous mutation in patient B (see Table 1). Because patient A had been deceased just the correspondent regions were mapped again. The variant c.2048_2052dupTAATA; p.Glu685* was confirmed with Sanger sequencing because 2 reads of the variant with next generation sequencing were not sufficient.
By WES, the compound heterozygous mutations c.230A > G and c.2048_2052dup in the gene coding for TBCD were found in both patients. The mother is a heterozygous carrier of the missense mutation. The father is a heterozygous carrier of the variant c.2048_2052dup TAATA; p.Glu685* in exon 24, which results in a premature stop codon and probably in the following to a nonsense-mediated messenger RNA decay. This could clearly be visualized by a significantly reduced signal of the affected allele in the cDNA (panel B) in comparison with wild type cDNA (panel A) (see Figure 4).
(A) Shows a healthy control, whereas the sequencing of exon 24 in the patient's complementary DNA. (B) Shows a decreased signal in the region of the duplicated TAATA-sequence (underlined).
The variant c.230A > G is predicted to be “deleterious” by PROVEAN, 13 “likely pathogenic” by ClinVar (submitted 2018, submission number: SCV000892499) 14 and “damaging” by sorting intolerant from tolerant (SIFT). 15 Furthermore, both variants are predicted to be “disease causing” by mutation taster. 16 In the following, we present a table with the known disease-causing mutations and references leading to a similar clinical presentation (see Table 2).
Known Disease-Causing Mutations in Tubulin Folding Cofactor D.
Discussion
We presented two patients with a severe neurodegenerative disease with 1 novel mutation in the TBCD gene and a particularly severe clinical phenotype.
The laboratory findings showed in both patients intermittent neutropenia. A possible explanation for the neutropenia could be the enhanced stability of the microtubules in TBCD deficiency, which was investigated and determined by Flex et al. 8 This has also been observed as a side effect of the drug taxol. Taxol stabilizes microtubules (Abal et al 27 ) and often leads to neutropenia (Rowinsky et al 28 ).
We contemplate a hormonal involvement of the disease because our female patient presented with premature thelarche at the age of 19 months. A TBCD-deficient patient with hirsutism had been described in the literature, 26 as well as a patient with cryptorchidism and hypothyroidism. 18 Moreover, our male patient had undescended testicles, which could be caused by multiple factors. Further investigation of TBCD deficiency would be necessary to clarify a hormonal interaction of the disease.
By reconstructing the diagnostic way of the 2 patients, we want to emphasize the importance of reevaluation and potential repetition of diagnostics, especially if new technology is available. The average coverage had improved remarkably while the amount of required material decreased in the last decade which allows us to analyze more samples per run. The percentage of target sequences covered at least 20-fold had increased just marginally from ∼97% to 98% in the time between the first and the last successful sequencing. This overall improvement, however, also included significant relative improvements of many outliers (see Figure 5).
This chart illustrates the increase of covered regions exceeding 20-fold with WES (CDS, and HGMD). The higher the coverage of coding sequences and disease-causing genes documented by the HGMD, the higher the probability to identify rare disease-causing mutations and interpret them correctly. While the coverage improved immensely in the last decade, the required input material decreased due to technological advance (CeGaT). The combined effect of the determinants facilitated diagnosing the TBCD deficiency.
Additionally, we can report a novel disease-causing mutation (c.2048_2052dupTAATA; p.Glu685*) that had not been described previously. The heterozygous mutation c.230A > G had been described by Tian et al 19 with a mild clinical manifestation. The variant c.230A > G in combination with the deletions of exons 28-39 leads to a heavy clinical manifestation with a phenotype similar to our patients (Zhang et al 17 ), as well as the homozygous mutation (Elmas et al 18 ). We compare the clinical presentation of the patients carrying the same variant c.230A > G in Table 3.
Abbreviations: EEG, electroencephalogram; MRI, magnetic resonance imaging.
The characteristics of phenotypes most likely depend on the remaining protein function and especially on the ability to interact with the cofactors and β-tubulin(Miyake et al 9 ).
The milder mutation in our case might be the missense mutation c.230A > G and determines the clinical picture. It remains unclear whether modulating factors are present in the different patients.
At first glance, the second TBCD allele does not explain the clinical difference, as in all 3 cases a loss of function is predicted. A dominant effect, eg by incorporation of some truncated nonfunctional forms into the DEα-/β-tubulin/C-super complex, is also unlikely, since the parents were completely normal. However, it cannot be excluded that the mutation c.907C > T (p.Arg303*) reported by Tian et al 19 is a hypomorphic nonsense mutation providing a relevant residual function. This may occur through mutation-induced alternative splicing with new splice sites or by exon skipping, leaving a protein with a significant remaining function (eg Roosing et al 29 ) or by a significant percentage of read through. This could solve the paradox. In the patient described by Tian et al, 19 the mutation c.230A > G would be the more severe defect than the hypomorphic stop, which would explain the mild phenotype in their patient(s) (Table 3). In contrast, the other patients in Table 3 with the heterozygous mutation likely would have a true stop as the second allele explaining their more severe phenotype. The homozygous missense mutation also leads to a severe phenotype because the residual protein function does not suffice.
The patient's complete genetic background should not be overlooked and be considered as a significant influence on the manifestation of the disease.
Forty-three patients with various mutations in the TBCD gene but different phenotypes had been described so far. The majority of patients developed a severe neurodegenerative disorder with an early onset within the first year; very few patients with mild clinical features could be identified. Aside from a symptomatic antiepileptic pharmacotherapy, there is no causal treatment until now. Most of the children developed severe respiratory problems and deceased within months or a few years.
The disease is probably underdiagnosed and not very widely known. It took a long time until the children could be diagnosed correctly. Early WESs did not result in the correct diagnosis, mostly due to a low coverage in the correspondent regions. The first description of the disease had not been published at that point. WES experienced a rapid development in the last decade due to the introduction of new diagnostic tools and efforts to increase efficiency. This progress was accompanied by an improvement of the coverage, which is an essential marker for the quality of a sample (Matthijs et al 30 ). Overall, this advance and the publications about the candidate gene were crucial for the diagnosis of the patients presented above.
The current WES has a better coverage and makes faster diagnosis possible. In general, unresolved cases may profit from the repetition of sequencing using improved processes that facilitate the identification of rare diseases and may therefore provide the basis for any development of a specific therapeutic approach.
Footnotes
Acknowledgments
The work was performed in the University Children's Hospital in Muenster, Germany.
Author Contributions
Christina M Quitmann drafted the manuscript and contributed to analysis and interpretation of data. A substantial contribution to the analysis and interpretation of data was made by Stephan Rust, Dr. rer. nat. and Janine Reunert, Dr. Biskup Dr. rer. nat., Dr. med. and Barbara Fiedler Dr. med. contributed to acquisition of data and interpretation. Thorsten Marquardt Prof., Dr. med. supervised the work and contributed to conception and interpretation of the data. All authors had critically revised the manuscript, finally approved it and agree to be accountable for all aspects of the work.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Ethical Approval
Not applicable, because this article does not contain any studies with human or animal subjects.