Identification of LAMA1 mutations ends diagnostic odyssey and has prognostic implications for patients with presumed Joubert syndrome

Abstract Paediatric neurology syndromes are a broad and complex group of conditions with a large spectrum of clinical phenotypes. Joubert syndrome is a genetically heterogeneous neurological ciliopathy syndrome with molar tooth sign as the neuroimaging hallmark. We reviewed the clinical, radiological and genetic data for several families with a clinical diagnosis of Joubert syndrome but negative genetic analysis. We detected biallelic pathogenic variants in LAMA1, including novel alleles, in each of the four cases we report, thereby establishing a firm diagnosis of Poretti–Boltshauser syndrome. Analysis of brain MRI revealed cerebellar dysplasia and cerebellar cysts, associated with Poretti–Boltshauser syndrome and the absence of typical molar tooth signs. Using large UK patient cohorts, the relative prevalence of Joubert syndrome as a cause of intellectual disability was 0.2% and of Poretti–Boltshauser syndrome was 0.02%. We conclude that children with congenital brain disorders that mimic Joubert syndrome may have a delayed diagnosis due to poor recognition of key features on brain imaging and the lack of inclusion of LAMA1 on molecular genetic gene panels. We advocate the inclusion of LAMA1 genetic analysis on all intellectual disability and Joubert syndrome gene panels and promote a wider awareness of the clinical and radiological features of these syndromes.


Introduction
Poretti-Boltshauser syndrome (PTBHS) (OMIM #615960) is a rare neuro-ophthalmological disease with phenotypes that may include a non-progressive cerebellar ataxia, delayed motor and language development and intellectual disability. There are additional ophthalmological phenotypes associated with this condition that may include ocular motor apraxia (OMA), myopia, strabismus and retinal dystrophy. On brain imaging cerebellar dysplasia, cerebellar cysts and cerebellar vermis hypoplasia are typically seen. The syndrome was first described in 2014 where the features of seven children (from five unrelated families) were described. 1 Subsequently, the underlying molecular defect was shown to be biallelic truncating or splice site mutations in LAMA1 2 which encodes the Laminin alpha-1 protein. Since these initial reports of PTBHS phenotypes there has been only a handful of publications detailing additional cases. [3][4][5][6][7] This suggests either that the disease remains ultra-rare, or that cases of PTBHS are not being recognized and are being incorrectly labelled. Another rare condition affecting the cerebellum is Joubert syndrome, a syndrome characterized clinically by a wide array of features including hypotonia, abnormal breathing patterns, OMA, ataxia and intellectual disability. Additional features include retinal dystrophy, nephronophthisis, liver fibrosis and skeletal dysplasia. 8 The condition was described in 1969 by Marie Joubert 9 and subsequently several more families with episodic hyperpnoea, abnormal eye movements and ataxia were noted. 10 Joubert syndrome has an estimated prevalence of 1/55,000-1/200,000. [11][12][13] Brain MRI in Joubert syndrome typically reveals a molar tooth sign (MTS) secondary to a deep interpeduncular fossa and elongation of the superior cerebellar peduncles. 10 Joubert syndrome is clinically heterogeneous [13][14][15][16] and there are currently 38 known genetic causes of Joubert syndrome. 8 Next-generation sequencing approaches have over the past decade rapidly improved our ability to diagnose rare inherited diseases [16][17][18][19][20] and these have been applied to patients with suspected Joubert syndrome.
Here, we review the clinical and imaging details of a child with developmental delay and ataxia, who had been labelled as having Joubert syndrome. We were able to establish a diagnosis of PTBHS following a review of imaging and whole-exome sequencing (WES) approaches. This prompted us to identify another three cases where similar diagnostic pitfalls led to diagnostic delays. Genomics England 100,000 genomes project Whole-genome sequencing (WGS) was performed through the Genomics England 100,000 Genomes Project (GE100KGP). All participants in the 100,000 Genomes Project have provided written consent to provide access to their anonymized clinical and genomic data for research purposes (https://re.extge.co.uk/ovd/) (last accessed 30/03/2021).

Deciphering developmental disorders study
A total of 13,451 individuals with severe, undiagnosed developmental disorders were recruited from 24 regional genetics services within the United Kingdom National Health Service and the Republic of Ireland. Families gave informed consent to participate, and the study was approved by the UK Research Ethics Committee (10/H0305/83 granted by the Cambridge South Research Ethics Committee, and GEN/284/12 granted by the Republic of Ireland Research Ethics Committee). Details on sample collection and genetic analysis pipelines have been described previously. 21 Genetic variants and linked phenotypic data have been accessed via DECIPHER (https://www.deciphergenom ics.org/) (last access 30/03/21).

Variants validation by Sanger sequencing
Sanger sequencing was utilized to confirm variants and their segregation from both parents where DNA samples where available. PCR amplification was performed using Taq PCR master mix (Qiagen) kit, as per the manufacturer's instructions.

Data availability statement
The authors confirm that the data supporting the findings of this study are available within the article and its supplementary materials. Further phenotypic or sequencing data are available from the corresponding author (J.A.S.), upon reasonable request. Genetic variants and linked phenotypic data from the deciphering developmental disorders study can be accessed via DECIPHER (https:// www.deciphergenomics.org/). The novel genetic variants in LAMA1 have been submitted to ClinVar.

Patients characteristics
The index family proband (NCL_Q73) was a child of 3 years of age that had been labelled as Joubert syndrome given her clinical features of ataxia, OMA and a possible molar tooth sign detected on brain MRI imaging ( Table  1, Fig. 1). A previous genetic analysis using a next-generation sequencing 29 genes panel of Joubert syndrome genes was negative. Her history was reviewed as well as her brain imaging (Table 1) and features seemed to diverge from a typical Joubert syndrome presentation. In particular, the brain MRI scans showed evidence of cerebellar cysts (

Exome sequencing data
Following informed consent, duo WES was carried out for the proband and her mother in order to pursue a precise molecular genetic diagnosis. WES data was initially filtered for biallelic changes in 38 genes known to cause Joubert syndrome (Supplementary Table 3) and monoallelic changes in SUFU. This led to no obvious underlying genetic causes in this set of genes. Filtering of variants was adjusted to examine all rare variants within the exome dataset and this identified two rare and predicted pathogenic alleles in LAMA1 (Fig. 1, Table 2). Using DNA obtained from both parents and the proband we confirmed the variants in LAMA1 by Sanger sequencing as well as their segregation from each parent. Biallelic variants in LAMA1 cause PTBHS, which would fully account for the clinical presentation and phenotype. The identification of this index case of PTBHS and the diagnostic pitfalls leading to an erroneous clinical diagnosis of Joubert syndrome led us to search for other similar mislabelled cases.

Identification of additional patients with undiagnosed PTBHS
The 100,000 Genomes Project provides a rich source of WGS data on patients with rare disease phenotypes. 22 We searched the whole rare disease dataset (73,988 genomes) for pathogenic alleles in LAMA1 and identified 2 probands (GEL-01 and GEL-02) with rare, pathogenic biallelic changes in LAMA1 (Fig. 1, Tables 1 and 2, Supplementary Figs 5 and 6) and brain MRI phenotypes consistent with PTBHS ( Supplementary Figs 7 and 8). These alleles had been filtered out by the standard tiering tables and the patients' physicians had received an 'exit questionnaire' suggesting no causative mutations had been identified and that the patients remained genetically unsolved. The most likely reason for this is that these patients were recruited into phenotypic groups 'Congenital malformations caused by ciliopathies' (specific disease Joubert syndrome) for patient GEL-01 and 'Motor disorders of the CNS' (specific disease cerebellar hypoplasia) for patient GEL-02, neither of which contained LAMA1 as a causative gene in the virtual gene panel applied. We have subsequently added LAMA1 as a potential causative gene to these panels via the Genomics England PanelApp (https://panelapp.genomicsengland.co.uk) so that future cases will not be missed. Finally, we examined the DECIPHER database 23 (https://decipher.sanger.ac. uk/) for patients with LAMA1 mutations and found an additional UK family with a proband (NCL_Q98) who had  Fig. 5) and a clinical diagnosis was confirmed to be PTBHS. 24 Finally, we examined published cases of PTBHS in the literature to determine longer-term outcomes for patients with this condition, given that this is a frequently asked question by both physicians and affected family members (Supplementary Table 4). Data are limited but there is some evidence of extremely good outcomes including adult patients with normal IQ levels, attending college and higher education and running their own businesses. The caveat is that the majority of known patients with PTBHS are still children and the long-term outlook is not yet known.

Estimated prevalence of PTBHS and Joubert syndrome in UK patient cohorts
In order to assess the relative prevalence of molecular diagnoses of PTBHS and Joubert syndrome in patients with intellectual disabilities and developmental disorders, we searched the Genomics England (GEL) 100,000 Genomes Project rare disease dataset as well as deciphering developmental disorders study for cases solved with either one of the 38 known Joubert syndrome genes (Supplementary Table 3) or pathogenic variants in LAMA1. Out of 8459 probands recruited in GEL with human phenotype ontology terms 'intellectual disability' and/or 'developmental delay', 18 cases (0.2%) were molecularly solved for Joubert syndrome genes versus only two cases solved with LAMA1 variants (Fig. 2A). Similarly, among 14,000 individuals with developmental disorders recruited in the UK deciphering developmental disorders study, 26 (0.2%) were solved for a Joubert syndrome gene and three cases solved with LAMA1 variants (Fig. 2B). Molecular genetics approaches, unbiased for clinical or radiological assessments, indicate thus that PTBHS is a considerably rarer cause of developmental disorders and intellectual disability than the collective of genes causing Joubert syndrome, at least in these UK cohorts. On the other hand, the relative contribution of LAMA1 to developmental disorders is likely comparable to the more common among the Joubert syndrome genes, when considered individually (Fig. 2).

Discussion
Joubert syndrome and PTBHS can present in very similar ways in infancy, and this may cause diagnostic uncertainty and confusion (Supplementary Table 1, Fig. 3). Clearly, it is vital that at this time, when patients are 'deviating' from normal development, appropriate investigations are planned. For multisystem syndromes such as Joubert syndrome it is desirable for an early diagnosis to be made, for example before the onset of end-stage kidney disease. For a more limited condition, there may be less urgency, but both family members and physicians will appreciate a precise diagnostic label so appropriate management plans can be put in place and the diagnostic work-up stopped. This is also relevant when considering allocations of public healthcare resources, as a molecular diagnosis of PTBHS implies that extra-CNS manifestations are unlikely to occur and do not need to be regularly screened for, in contrast to Joubert syndrome. 8 The main clinical similarities between Joubert syndrome and PTBHS are delays in motor and speech development, and OMA in the vast majority (Supplementary Table 1 SUFU cause a similar clinical phenotype in infancy. 25 We emphasize that the clinical findings in infancy do overlap but later in childhood the syndromes of Joubert syndrome and PTBHS diverge.

). A recent observation is that autosomal dominant variant in
Regarding brain imaging, pattern recognition is required for the identification and distinction of the two syndromes. The radiologist needs high-definition images in order to make a correct diagnosis and needs to be  qualified in the interpretation of neurological conditions. The field of posterior fossa anomalies may not be an area of expertise for most radiologists. This point is important as the correct interpretation of brain imaging will lead to the correct set of genetic investigations to be performed and will result in a faster, more cost-effective diagnostic odyssey for the patient.
For Joubert syndrome, the MRI imaging hallmark is the molar tooth sign (MTS), but in practice, there is an imaging spectrum of additional infra-and supratentorial anomalies. 26 It is noteworthy that the MTS is variable in peculiarity-from very pronounced to less marked, in the literature also called 'mild MTS'. Non-specialists may easily overlook mild MTS features on brain MRI. The findings of mild MTS have often been reported in Joubert syndrome associated with mutations in NPHP1, C5orf42, FAM149B1, CBY1. 27 The mild MTS seen in some types of Joubert syndrome is not distinguishable from mild MTS in patients with SUFU variants, emphasizing the role of targeted molecular genetic investigations for these conditions. The differential diagnosis for MTS includes pontine cap dysplasia, but the clinical context and remainder of imaging are very different. 28 We would like to emphasize that there is no imaging overlap between Joubert syndrome and PTBHS (Supplementary Table 2). In PTBHS brain imaging there is a 'hierarchy' of findings, of which cerebellar dysplasia is the leading sign, followed by cerebellar cysts. The differential diagnosis of cerebellar dysplasia and cysts includes some severe forms of congenital muscular dystrophies however, other clinical contexts and additional MRI anomalies do not allow for diagnostic confusion. The presence of a typical MTS has a strong impact on the presumed diagnosis and the further workup. Given these implications, the clinician that requested MRI images should carefully review the imaging findings and act as a 'gate-keeper' for a premature label of MTS and Joubert syndrome.
The accessibility of WGS and WES in both research and clinical care settings has allowed huge advances in our understanding and diagnosis of rare diseases. 22,29 Indeed, genetic work-up has become standard of care for several neurodevelopmental disorders and the diagnostic yields by WES in patients with an intellectual disability or global developmental delay is estimated around 35%. 30 The testing modalities (sequencing, array-based techniques, karyotyping) and scope of genetic testing (selected gene panel, WES or WGS) depend on the clinical presentation and recommendations are likely to evolve with technological development. Emerging evidence suggests that WES may be preferable over a targeted analysis of selected genes and that WES also outperforms chromosomal microarrays as an initial testing approach in neurodevelopmental disease. 31 Although overall coverage is usually better with targeted analyses, WES was shown to cover >98% of mutations identified on targeted next-generation sequencing panels. 32 Interestingly, a simulation study that asked clinicians to choose a commercially available gene panel as an alternative when ordering singleton WES for the various suspected monogenic diseases found that 23% of patients would not have been solved had this panel been applied instead of WES. 33 What is more, panel analyses are quickly outdated as novel genetic aetiologies for disease emerge and gene discovery in the arena of neurodevelopmental disorders is fast moving. 20,27 As an example, 11% of patients negative on a comprehensive 'Mendeliome' panel received a diagnosis via WES, simply because the panel did not include the most recent gene-disease relationships identified after the panel was established. 34 Finally, WES is cost-efficient, with a considerable decrease in healthcare expenditures and downstream medical interventions once WES was performed and WES also compared favourably with targeted analysis, especially if a negative panel analysis leads to further investigations and subsequent re-analysis by WES or WGS. 33,35 Indeed, the cases that we report here corroborate the notion that negative findings in a selected gene panel analysis should be followed-up by an unbiased investigation using WES or WGS. Taken together, the limitations of panel analyses are not of technical nature; on the contrary, targeted analyses usually provide increased read depth, sequence coverage and detection of exon-level deletions and/or duplications. The examples above as well as the cases reported here rather highlight the difficulties for an a priori assignment of potential genetic causes to a phenotype. This is an important limitation for gene panel analysis, where precise phenotypic characterization is required and dictates the set of genes that are investigated. Virtual gene panels, where untargeted WES or WGS is performed, and variant interpretation then restricted to a selected number of genes have several advantages: generated data is easier to manage, secondary findings are unlikely to occur, but the scope of analysis can be extended if no pathogenic variants are detected, or new genes are shown to cause the investigated phenotype. We have shown here that there is a certain phenotypic overlap, especially in infancy, between Joubert syndrome and PTBHS, and therefore it seems a pragmatic approach to add the single LAMA1 gene to Joubert syndrome gene panels so this alternate diagnosis can be detected. In addition, efforts to re-analyse unsolved Joubert syndrome-like patients and those with cerebellar dysplasia and cysts should be made by opening up virtual gene panels to other genetic phenocopies, such as LAMA1. Expert multidisciplinary teams can be utilized to update and extend virtual gene panels and report back the alleles and their pathogenicity in context with the clinical phenotype, which can evolve as the child develops.
In discovering these cases of PTBHS, we were struck by the profound impact the genetic diagnosis made on each of the families. In all four cases, there was a delay caused by incomplete genetic analysis and failure to include LAMA1 in the gene panel applied. Due to this lack of molecular genetic diagnosis, in all four cases, the clinical diagnosis defaulted to Joubert syndrome. The molecular genetic diagnosis of LAMA1 mutations led to a new diagnostic label and a feeling of loss and insecurity for the families involved. A new diagnosis also has some prognostic implications. Compared to the multisystem features of Joubert syndrome, PTBHS has a much more limited phenotype. To date, no other organs aside from the brain are involved and there is no evidence for disease progression. In particular, it is not associated with the renal and liver problems which may be seen in Joubert syndrome and individuals with an established diagnosis of PTBHS do not require regular monitoring for these. Although OMA persists into adult life. With that in mind, we tried to draw some conclusions from the older patients that have been reported in the literature. This confirmed the non-progressive nature of the PTBHS phenotype and overall favourable outcomes including some examples of normal intellectual ability and independence but we suspect this data is both incomplete and subject to bias. In line with a somewhat milder phenotype, we also detected a 50-year-old individual in gnomAD carrying a homozygous predicted loss-of-function variant in LAMA1 (c.858þ1G>T).
Next-generation sequencing allows an unbiased analysis of the whole exome or genome and has become more widely available for use as a first-line diagnostic tool for the investigation of patients with neurodevelopmental disease, including cerebellar disorders. It has the potential to resolve those cases with suspected Joubert syndrome and disorders such as PTBHS which may clinically mimic Joubert syndrome in infancy. 36 We present four realworld examples highlighting the strength of an unbiased genetic approach over a targeted panel analysis and of how a precise molecular diagnosis allows some clarity in terms of clinical monitoring for the physician and potential long-term outcomes for the patient.

Supplementary material
Supplementary material is available at Brain Communications online. Abbreviated summary; Powell, Olinger et al. report four gene panelnegative kindreds with presumed Joubert syndrome where critical review of clinico-radiological data coupled with exome or genome sequencing re-assigned a corrected diagnosis of Poretti-Boltshauser syndrome with prognostic implications. Phenotypic overlap and limited disease hallmark recognition advocate for broad, unbiased genetic testing in rare neurodevelopmental disorders.