Introduction

NDDs are a clinically and genetically heterogeneous group of prenatal or neonatal onset pediatric conditions that impair the brain’s structure and/or function, typically resulting from disruption of tightly regulated developmental processes (Khan et al. 2019; Mitani et al. 2021). NDDs arise from the disruption of crucial neurodevelopmental mechanisms or pathways and are associated frequently with abnormalities in neurogenesis, glial and neuronal proliferation, neuronal migration, synapse formation, and myelination (Homberg et al. 2016). Subsequent impaired cognitive function may affect behavior, motor skills, learning, reasoning, memory, problem-solving, decision-making, communication, and attention, culminating in cognitive, communicative, behavioral, and psychomotor impairments (Bogaert et al. 2023; Cardoso et al. 2019; Morris-Rosendahl and Crocq 2020). Intellectual disability (ID) in mild to severe forms are the most predominant types of NDD and affects 1–3% of the world’s population, while its prevalence is twofold in the developing world (Riazuddin et al. 2017). Common manifestations of NDDs include developmental delay (DD), specific learning and motor disorders, communication disorders, autism spectrum disorders, attention deficit hyperactivity disorder, movement disorders, and epilepsy (Morris-Rosendahl and Crocq 2020; Parenti et al. 2020). Up to 50% of NDD cases are estimated to have a genetic etiology (Servetti et al. 2021). In the last decade, the widespread integration of next-generation sequencing approaches, most notably ES, into the diagnostic process has increased the number of affected individuals who receive a molecular diagnosis (Farwell et al. 2015; Lee et al. 2014; Wright et al. 2023; Yang et al. 2013, 2014). With an average diagnostic yield of 30–50%, ES has emerged as a key diagnostic tool in clinical and research environments and is considered a first-tier (Clark et al. 2018; Srivastava et al. 2019) diagnostic approach. This approach is particularly helpful when investigating patient-parent trios or related affected or unaffected individuals, followed by confirmation and co-segregation analysis of candidate variants within the family through Sanger sequencing, further enhancing the diagnostic process. ES offers a relatively fast, inexpensive, and precise molecular diagnosis in affected individuals with overlapping, unexplained, heterogeneous disorders impacting the central and peripheral nervous system.

This study aimed to identify the genetic etiology of NDDs in four unrelated Pakistani families using ES as a first-tier diagnostic approach. We identified disease-causing variants in established NDD-associated genes in all four families, including one previously unreported variant in RELN and three recurrent variants in VPS13B, DEGS1, and SPG11.

Materials and methods

Research participants and sample collection

This study was approved by the Institutional Review Board and Ethical Committee at the National University of Medical Sciences, Rawalpindi, Pakistan, and the Department of Biological and Biomedical Sciences, The Aga Khan University, Karachi, Pakistan to ensure adherence to ethical guidelines. Four Pakistani families, including three consanguineous and one non-consanguineous, were ascertained, with all families having at least one member exhibiting NDD phenotypes likely of genetic origin. We obtained written informed consent from each proband or their legal guardians in accordance with the Declaration of Helsinki. Detailed clinical histories were taken for each affected individual and peripheral blood samples were collected from all consented affected and unaffected family members. Genomic DNA extraction from blood samples was performed using the GeneJET Genomic DNA Purification Kit (Thermo Scientific).

Exome sequencing, variant filtration, and in silico analyses

We performed paired-end ES using genomic DNA of at least one affected individual from each family. Genomic DNA was fragmented and target enriched for exome capture using the Twist Human Core Exome library preparation kit (Twist Biosciences) following the manufacturer’s protocol. The captured libraries were then sequenced on the Illumina Novaseq 6000 platform, generating 150 bp paired-end reads. This achieved a mean depth of 68–84× across individuals, with 87–90% coding region coverage at a minimum of 20 reads. The sequencing yielded FASTQ files that underwent downstream bioinformatics analysis for variant calling and filtration. No copy number variants analysis was performed in these families.

Quality control (QC) of raw paired-end ES reads in FASTQ format was first evaluated with FastQC. Based on the QC results, the low-quality reads and adapter sequences were trimmed (or filtered out) using the fastp tool (Chen et al. 2018). Trimmed and filtered reads underwent a second round of QC with FastQC to evaluate the effectiveness of fastp preprocessing. The trimmed and filtered reads were then aligned to the hg19 human reference genome with BWA-MEM (Li 2013) resulting in a sequence alignment/map (SAM) file. The aligned sequences were processed with Picard tools to generate a sorted and deduplicated binary alignment map (BAM) file. Germline single nucleotide variants (SNVs) and indels were called and processed with the Genome Analysis Toolkit (GATK4) (DePristo et al. 2011; McKenna et al. 2010) following the GATK Best Practices workflow (Van der Auwera et al. 2013) resulting in a variant call format (VCF). Variants were annotated by wANNOVAR (Yang and Wang 2015) to add functional and interpretive information from databases including Gene Ontology, Genome Aggregation Database (gnomAD), Exome Aggregation Consortium (ExAC), 1000 Genomes, NHLBI-ESP, dbSNP, SIFT, PolyPhen-2, MutationTaster, and CADD as well as clinical associations from ClinVar and GWAS catalog.

Reads with a minimum of 10× coverage were retained for further analysis. Variants were filtered to remove reported SNVs with minor allele frequencies (MAFs) > 0.01 in public databases including dbSNP, 1000 Genomes, gnomAD, ExAC, and TOPMed BRAVO. Exonic or splice variants with MAFs < 0.01 consistent with the inheritance patterns of disease and previous association with NDDs were prioritized as candidates. These candidate variants underwent bidirectional Sanger sequencing validation and segregation analysis in DNA from all available consented family members using BigDye Terminator v3.1 on an ABI 3730 platform (List of primers is available in Supplementary Table 1). Bioinformatics pipeline results were validated using VarSome Clinical (v11.6). Variants were classified and interpreted according to the guidelines of the American College of Medical Genetics and Genomics (ACMG) (Richards et al. 2015).

Results

Family 1

Clinical findings

A consanguineous Pakistani family with two affected siblings 1:3 and 1:5 (Fig. 1) exhibited a severe neurodevelopmental condition characterized by ID, DD, speech and language delay, and microcephaly. Both affected individuals were delivered full-term by spontaneous vaginal delivery with no reported pregnancy complications. Individual 1:3 (male, age 17 years at last clinical assessment) had seizure onset at 6 months after a fever, lasting 1 year and requiring hospitalization. Hands and feet contractures were noticed at around 6 months of age. He has microcephaly (−5.44 SD), severe ID, DD, speech delay, musculoskeletal anomalies, muscle spasticity, hypotonia, hand tremors, short stature, bed-confined, inability to self-feed, crawl, and stand with support by age 3 years, and is only able to speak a few words by the age of 16 years. No myopia or retinopathy was observed in this patient. The individual exhibited sociable behavior, typical facial dysmorphism, and joint hyperextensibility. Sensation was normal in his right toe but absent in his left toe. He tried to respond to words but was unable to speak. Additionally, he had a disturbed sleeping pattern. Individual 1:5 (female, age 20 years at last clinical assessment) has microcephaly (−4.10 SD), severe ID, DD, and speech delay. By the age of 5 years, she was only able to speak 1–3 words. She exhibited sociable and cheerful behavior, although occasional episodes of aggression were observed. She had low vision, night blindness, joint hypermobility, and typical facial characteristics. She had a normal sensation in her toes. She tried to respond to some words but was unable to speak. She demonstrated independent self-feeding and eating skills. Detailed clinical features are described in Supplementary Table 2.

Fig. 1
figure 1

Pedigrees and genetic findings in families with NDDs. Top bold text shows the name and identified variant in each family. The double horizontal lines in pedigrees indicate consanguinity; squares and circles symbolize males and females, respectively. Filled symbols indicate affected individuals; unfilled symbols indicate healthy individuals; diagonal line, deceased. The genotypes for identified variants are indicated below each symbol where DNA was available. Electropherograms of affected (upper panel) and healthy (lower panel) individuals below each pedigree highlight the precise location (dotted rectangle) of variants in the DNA sequence. Asterisks in the pedigree indicate family members whose DNA was unavailable. Ref, reference allele; Alt, alternative allele

Full size image

Genetic findings

ES in both affected siblings (1:3 and 1:5) and subsequent bioinformatic filtering identified ten variants (Supplementary Table 3) including a homozygous variant (NM_152564.5):c.11758C>T (p.Arg3920Ter) in VPS13B shared by both affected individuals (Fig. 1; Table 1). The identified variant is rare in public databases (MAF of 0.000004957 in gnomAD v4.0.0). We prioritized the VPS13B variant due to previous association of the gene with autosomal recessive Cohen syndrome (Kolehmainen et al. 2003) overlapping with the clinical presentation of affected individuals in Family 1. Bi-directional Sanger sequencing validated the segregation of the VPS13B variant with disease status across the family. This variant has not been reported in the literature although ClinVar submissions (VCV000550808.37) classifies this variant as pathogenic/likely pathogenic with no information available on phenotype.

Table 1 List of potential disease-causing variants identified by ES in this study
Full size table

Family 2

Clinical findings

A consanguineous Pakistani family with one affected female individual (2:3) aged two years (at last clinical assessment) was ascertained from the Sindh province (Pakistan). She was born at full term through a C-section. She has shown delayed developmental milestones, including delays in sitting, babbling, walking, and learning difficulties. She has strabismus in both eyes, stiffness in both the upper and lower limbs, and muscle atrophy in both limbs. She has been assessed clinically for her aggressive behavior. Over time, the disease symptoms have exhibited a progressive and deteriorating trend. She experiences recurring seizures, typically once a month. Presently, she faces challenges in thriving, is confined to a bedridden state, and necessitates assistance for fundamental self-care activities and feeding.

Genetic findings

In Family 2 proband-only ES identified multiple variants (Supplementary Table 3) including a previously unreported homozygous nonsense variant in RELN (NM_005045.4):c.4696C>T (p.Gln1566Ter) that segregated with disease in the trio (Fig. 1; Table 1). RELN is associated with autosomal recessive Lissencephaly 2 (Norman-Roberts type; OMIM 257320) (Hong et al. 2000), and lateral temporal lobe epilepsy-7 (OMIM 616436) (Dazzo et al. 2015). Lissencephaly, sometimes known as “smooth brain” is a severe neurodevelopmental condition in which neuronal migration is impeded, resulting in a thicker cerebral cortex with a smoothed-out, simplified fold pattern. Lissencephaly is characterized by significant hippocampal and cerebellar abnormalities (Hong et al. 2000). Due to previous association of RELN variants with the disease and its concordance with the proband’s clinical presentation, we prioritized this variant. The identified variant is hitherto unreported and is absent in gnomAD v4.0.0 and in our in-house healthy control database, and it is classified as Likely pathogenic (Table 1).

Family 3

Clinical findings

A consanguineous Pakistani family with second-cousin parents has three affected members (Fig. 1) exhibiting a severe NDD marked by global developmental delay (GDD), speech delay, hypotonia, and microcephaly. Proband 3:3 was a 7-year-old (at last clinical assessment) female who presented GDD, ID, severe microcephaly (−6.05 SD), spasticity, hypotonia, lack of speech, motor delay, inability to sit/walk/self-feed, and limb contractures. Her mother reported she did not cry at birth and exhibited transient neonatal hyperkinesia that later resolved. She was born full-term through a cesarean section. Her brain magnetic resonance imaging (MRI) noted moderate to severe brain atrophy. Proband 3:4 was a 14-year-old (at last clinical assessment) female who exhibited GDD, ID, microcephaly (−6.17 SD), spasticity, hypotonia, lack of expressive speech, motor delay, limb contractures, and scoliosis. She was born full-term through spontaneous vaginal delivery. Proband 3:5 was a deceased 9-year-old (at last clinical assessment) male who presented with GDD, ID, hypotonia, lack of speech, inability to sit/walk/self-feed, childhood milestone regression, and limb contractures. He was also born full-term via spontaneous vaginal delivery.

Genetic findings

Trio-ES of an affected individual (3:3) and both parents (3:1 and 3:2) identified a homozygous recurrent variant (NM_003676.4):c.517C>T (p.Arg173Ter) in DEGS1 (Fig. 1; Table 1). The variant causes a premature translational stop signal (p.Arg173Ter) and is predicted to result in nonsense-mediated decay (NMD). DEGS1 is associated with autosomal recessive leukodystrophy, hypomyelinating, 18 (HLD18, OMIM 618404). HLD18 is hallmarked by an onset of GDD in early infancy. In the most severe cases, affected individuals have impaired psychomotor development, which includes the inability to sit or walk independently, aphasia or poor speech, dystonia, spasticity, and seizures may occur in some cases (Pant et al. 2019).

Segregation analysis showed that the DEGS1 variant segregates in a recessive inheritance pattern in the family, homozygous in affected individuals, and heterozygous in unaffected parents. A complete list of additional variants identified after bioinformatic filtering is provided in Supplementary Table 3.

Family 4

Clinical findings

A 19-year-old male proband (4:3) from a non-consanguineous pedigree (Fig. 1) presented with a progressive learning disability, task-specific tremors, and gait abnormality beginning in late childhood. He achieved normal early developmental milestones. At the age of 12 years, intention tremors started in his right arm before progressing to involve the left arm bilaterally by the age of 15 years. He also developed progressive gait abnormalities and leg spasticity without lower limb tremors. Cognitive decline became evident impairing his academic performance. Clinical exam showed bilateral arm tremors on intent, leg spasticity, and normal muscle strength. Spiral drawing testing supported essential tremor. He has four unaffected siblings and has no family history of a similar neurological condition. Brain MRI at age of 16 years and nerve conduction studies and electromyography at age 18 years were normal. Extensive biochemical testing, including thyroid function, copper studies, ceruloplasmin, and alpha-fetoprotein were unremarkable. Despite a 7-year long clinical testing history, no definitive diagnosis was achieved by the age of 19 years.

Genetic findings

Parent-proband Trio-ES identified a recurrent homozygous frameshift deletion in SPG11 (NM_025137.4):c.5769del (p.Ser1923ArgfsTer28) (Fig. 1; Table 1) and a missense variant in PRKAG3 (Supplementary Table 3). Loss of SPG11 function is an established disease mechanism in autosomal recessive SPG11-related neurological conditions (Stevanin et al. 2007) overlapping with the phenotypes of proband. Hereditary spastic paraplegias (HSPs) are a heterogeneous group of neurodegenerative disorders characterized by progressive weakness and spasticity of the lower limbs associated with corticospinal axon degeneration.

The SPG11 variant has been reported previously as pathogenic in individuals with HSP (Kara et al. 2016; Paisan-Ruiz et al. 2008; Riazuddin et al. 2017; Schneider et al. 2012; Stevanin et al. 2008; Wakil et al. 2012; Zulfiqar et al. 2019). The variant segregates with disease in the pedigree; it is homozygous in the affected individual, heterozygous in the unaffected parents, and heterozygous or wild type in siblings (Fig. 1).

Discussion

Identification of a disease-causing gene(s) is a key step towards understanding the pathophysiological mechanisms of NDDs which enables precise molecular diagnosis, genetic counseling about recurrence risk, prenatal testing, disease management, and potential development of personalized treatments (Blesson and Cohen 2020; Parenti et al. 2020; Wright et al. 2018). Furthermore, definitive molecular diagnosis may prevent unnecessary testing, provide prognostic information, facilitate access to support services, end the diagnostic odyssey, and improve quality of life by providing psychological benefits (Carter et al. 2023). Despite advancements in diagnostic genetic testing, such as ES, achieving a precise molecular diagnosis for NDDs can be challenging due to diverse genetic underpinnings and clinical variability, where more than half of individuals may not get a definitive molecular diagnosis (Wojcik et al. 2023), face misdiagnosis, or receive multiple diagnoses, leading to a diagnostic odyssey.

ES has become a common clinical diagnostic tool for NDDs, with a reported diagnostic yield of ~30–50% (Clark et al. 2018; Srivastava et al. 2019), given its ability to provide a rapid, cost-effective, and highly accurate molecular diagnosis for patients with overlapping, unexplained, heterogeneous NDDs, we utilized ES in four unrelated NDD families for genetic diagnosis.

In Family 1, we identified a rare homozygous variant VPS13B (NM_152564.5):c.11758C>T (p.Arg3920Ter) as a likely disease cause. Pathogenic variants in VPS13B that segregate in an autosomal recessive inheritance pattern are associated with Cohen syndrome (El Chehadeh et al. 2010; Kolehmainen et al. 2003). There was marked intra-familial phenotypic variability among both affected individuals. Overlapping clinical characteristics among both affected individuals, such as ID, DD, speech delay, and microcephaly are consistent with Cohen syndrome.

VPS13B is a peripheral Golgi membrane protein that plays a crucial role in important cellular processes such as intracellular protein transport, sorting, vesicle trafficking, and maintaining Golgi integrity. It possesses functional motifs and putative transmembrane domains (Duplomb et al. 2014; Kolehmainen et al. 2003; Seifert et al. 2011). The C terminus of VPS13B localizes protein to the Golgi apparatus (Seifert et al. 2011). The identified variant p.Arg3920Ter introduces a premature stop codon in the C-terminus, resulting in a truncated protein lacking the final 78 amino acids that encompass the Golgi localization domain. Previous reports have identified disease-causing variants p.Ser3945GlnfsTer22, p.Asn3954Lysfs*60, p.Pro3944LeufsTer41 and (NM_152564.5):c.6657+1G>A, farther towards the C-terminus (El Chehadeh et al. 2010; Kolehmainen et al. 2004). Disruption of C-terminus may lead to defects in proper subcellular targeting of VPS13B to the Golgi apparatus (Seifert et al. 2011). This results in pathogenic defects in Golgi structure and function. Therefore, truncation of the C-terminal domain provides a molecular mechanism linking the p.Arg3920Ter variant to pathogenesis associated with Golgi dysfunction.

In family 2, we identified a hitherto unreported homozygous nonsense variant p.Gln1566Ter in RELN, leading to a premature stop codon that likely results in loss of Reelin protein. Reelin is an essential protein for brain growth and physiological activities. Functional analysis in a reeler loss of function mouse model reported that RELN leads to neurological symptoms with main features including ataxia, tremors, and a distinctive ‘reeling’ gait (Di Donato et al. 2022). Furthermore, multiple genome-wide association studies have also established a genetic link between RELN and various psychiatric disorders, including schizophrenia and autism spectrum disorders (Ishii et al. 2016). Biallelic variants in RELN are associated with a form of lissencephaly with cerebellar hypoplasia (Di Donato et al. 2022).

In Family 3, we identified a recurrent pathogenic variant (p.Arg173Ter) in DEGS1 associated previously with autosomal recessive HLD18 (Dolgin et al. 2019; Pant et al. 2019). DEGS1 encodes a Δ4-dihydroceramide desaturase, an endoplasmic reticulum lipid desaturase that catalyzes the final step in the de novo ceramide (Cer) biosynthesis pathway. Pathogenic variants in DEGS1 lead to dihydroceramide (DhCer) accumulation and DhCer/Cer imbalance which may be toxic for myelinating oligodendrocytes and disruptive for myelin biogenesis and neuronal health (Pant et al. 2019). Loss-of-DEGS1 function is a known mechanism of disease (Pant et al. 2019). The phenotypes of our patients are consistent with those reported previously, adding further evidence to the pathogenicity of the DEGS1 variants.

Family 4 exemplifies a diagnostic odyssey resolved by genetic diagnosis. In this family a recurrent homozygous SPG11 (NM_025137.4):c.5769del (p.Ser1923ArgfsTer28) variant segregated with disease. SPG11 is associated with autosomal recessive Spastic paraplegia 11 (OMIM 604360), caused by loss of function mechanism. SPG11 encodes spatacsin, a protein with a role in the maintenance and/or growth of neuronal axons and intracellular cargo trafficking (Pérez-Brangulí et al. 2014). The identified homozygous variant was reported previously as disease-causing in individuals with HSP (Kara et al. 2016; Paisan-Ruiz et al. 2008; Riazuddin et al. 2017; Schneider et al. 2012; Stevanin et al. 2008; Wakil et al. 2012; Zulfiqar et al. 2019) and clinical presentation of our patient was concordant.

In conclusion, we report four pedigrees with likely disease-causing variants in established NDD-associated genes, facilitated by large family structures ascertained in this study. This study demonstrates the utility of ES as a diagnostic tool for heterogeneous neurological disorders. Our study expands the clinical and mutational spectrum of NDDs.

Web resources