Rett Syndrome Diagnostic Odyssey: Limitations of NextGen Sequencing

Abstract

Typical (or Classic) Rett syndrome (RTT) is an X-linked neurodevelopmental disorder characterized by a period of regression, partial or complete loss of purposeful hand movements and acquired speech, impaired gait, and stereotyped hand movements. In over 95% of Typical RTT, a pathogenic variant is found in the Methyl-CPG binding protein 2 gene (MECP2). Here, we describe a young woman with clinically diagnosed Typical RTT syndrome who lacked a genetic diagnosis despite 20 years of investigation and multiple rounds of sequencing the MECP2 gene. Recently, additional genetic testing using next-generation sequencing was completed which revealed a partial insertion of the BCL11A gene within exon 4 of MECP2, resulting in a small deletion in MECP2, causing likely disruption of MeCP2 function due to a frameshift. This case demonstrates the ever-changing limitations of genetic testing, as well as the importance of continual pursuit of a diagnosis as technologies improve and are more widely utilized.

I. Introduction

Rett syndrome (RTT) is an X-linked neurodevelopmental disorder primarily caused by pathogenic variants resulting in loss of function of the Methyl-CPG binding protein 2 gene (MECP2).(Amir et al., 1999) The disorder was first recognized by Andreas Rett in 1966,(Rett, 1966) further corroborated by Hagberg in 1983, describing the condition for the first time as RTT.(Hagberg et al., 1983) Diagnostic criteria for typical RTT and atypical RTT were defined most recently in 2010, requiring individuals to have a period of regression followed by stabilization, and to meet 4/4 main criteria for typical RTT, or 2/4 main criteria and 5 supportive criteria for atypical RTT.(Neul et al., 2010) Currently, in >95% of individuals with typical RTT, a pathogenic variant is identified in the MECP2 gene.(Neul et al., 2008; Zahorakova et al., 2016) We describe a young woman diagnosed with typical RTT by clinical criteria without a genetic diagnosis until age 36, despite several rounds of genetic testing that included various methodologies of sequencing MECP2. Finally, in 2021, additional evaluation using a genetic panel identified a partial insertion of the BCL11A gene within exon 4 of MECP2, resulting in MECP2 loss of function. This case illustrates the evolving landscape of genetic testing technologies and the limitations of Next Generation (NextGen) sequencing due to deletions, duplications and insertions. (Singh et al., 2021) It also highlights the importance of clinical phenotyping and standard clinical criteria, as genetic testing continued to be pursued due at least in part to her meeting criteria for typical RTT. Without meeting the defined consensus clinical criteria for RTT, the underlying genetic etiology may not have been ultimately discovered.

II. Clinical Case Report

Clinical data was collected under IRB approval (COMIRB 13-2020). A 27-year-old woman was evaluated in the Rett Clinic at Children’s Hospital Colorado in 2013 given clinical concerns for RTT. She had already had extensive genetic testing at that time which was negative (Figure 1). She was given a diagnosis of RTT using established clinical criteria, and had a Rett Clinical Global Impression of Severity (CGI-S) score of 4 – Moderately Impaired (Neul et al., 2015). She met all criteria for Typical RTT: a period of regression (around 18-24 months of age), loss of acquired purposeful hand skills (at 18 months lost ability to self-feed and grab a bottle), loss of acquired speech (lost acquired single words around 18-24 months), gait abnormalities (wide based gait with limited gait distance), hand stereotypies (mouthing hand movements). Epilepsy was formally diagnosed at 3 years of age with electrographic seizures that lacked clinical correlate. Clinical seizures began around age 12 and were characterized by eyes rolling back, arms and legs stiff/flexed, jaw clenched, lasting approximately 30 seconds with a post-ictal period. These seizures were defined as focal motor seizures with impaired awareness, and lamotrigine was initiated to treat these seizures.

Figure 1.

Timeline of genetic testing.

Genetic evaluation was initially performed at Baylor College of Medicine in 2005 utilizing Sanger sequencing of the MECP2 gene, with no pathological variants in MECP2 identified at that time. Additional expanded genetic testing for Rett-like conditions using a NextGen sequencing panel was performed by Greenwood Genetics Clinic in 2013, which reported no identified pathogenic variants in ARX, FOXG1, PNKP, ATRX, MBD5 , SLC2A1, CDKL5, MECP2, SLC9A6, CNTNAP2, MEF2C, TCF4, EHMT1, OPHN1, UBE3A, FOLR1, PCDH19, ZEB2 and NRXN1. A chromosomal microarray was performed by Colorado Genetics Lab in 2013, which demonstrated a 522 kb, paternally inherited duplication in Xq27.3, not containing the MECP2 locus, and assessed as non-contributory to the clinical condition given the paternal inheritance from an unaffected father. Despite the lack of information provided from these genetic tests, the parents and clinical team felt that a genetic etiology underlined the clinical condition and remained determined to pursue additional genetic testing to provide both information regarding prognosis and identify potential research opportunities.

In November 2021, an Autism/Intellectual Disability Xpanded panel from GeneDx was completed at the request of the University of Colorado Neurogenetics Clinic, which identified a genomic rearrangement between MECP2 and BCL11A. This rearrangement consisted of a partial duplication of intron 1, exon 2, and intron 2 of the BCL11A gene (Chr2:60702464-60777252, 74kb) inserted into the coding sequence of exon 4 of MECP2, and a partial deletion of MECP2 exon 4 coding sequence. This rearrangement results in a frameshift variant in MECP2 predicted to result in a protein truncation (c.1156_1194delins(74789), p.Leu386fs NM_004992.3) loss of function (Kim et al., 2006; Krishnaraj et al., 2017). This partial duplication of BCL11A is about the same size as the whole MECP2 gene. MECP2 has 2 main transcripts due to alternative splicing. However, all transcripts retain the last exon, which is the largest. The MANE (Matched Annotations from NCBI and EBI)(Morales et al., 2022) transcripts identify this last exon as exon 4. It is predicted that this duplication would interrupt all transcripts of this allele. See Figure 2 for illustration of duplication/deletion.

Figure 2:

BCL11A duplication and MECP2 deletion

At last visit in Rett clinic in August of 2022 she continued to be moderately impaired, with a CGI-S of 4. She was non-verbal and did not communicate with an augmentative communication device but had good eye contact with the examiner (>30 seconds). She lacked any functional hand use and continued to have frequent disruptive hand stereotypies. She was able to ambulate independently, but she displayed a parkinsonian gait with difficulty initiating steps and a shuffling pace. She continued to have limited walk distance. This parkinsonian gait led to an initiation of carbidopa/levodopa but limited efficacy was observed over the next 4 months. She continued to have monthly seizures, treated with lamotrigine, and her family was satisfied with the seizure control provided by this treatment and no changes to seizure management were initiated.

IV. Discussion

This case of an affected individual meeting the clinical diagnosis of RTT but having repeated negative genetic testing results illustrates potential gaps in both Sanger and NextGen sequencing. However, timing and technology used for genetic testing is relevant. In this case, MECP2 was sequenced twice (2005-Sanger sequencing and 2013-NextGen sequencing) prior to the Autism/ID panel, and the 38bp deletion caused by the insertion was not reported. While recent sequencing did reveal the insertion/deletion, further examination is warranted to understand why this change would have been missed on prior testing. This case also underscores the importance of clinical phenotyping. Without the existence of clear criteria for Typical RTT and her meeting those criteria, a genetic diagnosis may not have been sought, or this important genetic result may have been overlooked. Given the clinical diagnosis of typical RTT and specific, newly FDA-approved treatments (Glaze et al., 2019; Harris, 2023; Neul et al., 2023) as well as the potential for gene-replacement approaches (NCT05898620, NCT05606614), pursuit of a precise genetic diagnosis can have tangible impact on medical decision making.

The initial MECP2 sequencing completed at Baylor was by Sanger Sequencing as Next Generation sequencing was in its inception.(Kchouk et al., 2017) Sanger sequencing should have resulted in a single sequence of the entire MECP2 gene, with longer reading frames when compared with NextGen sequencing. While accuracy of NextGen sequencing and Sanger sequencing is now thought to be comparable,(Beck et al., 2016) Sanger sequencing is still considered the gold standard and continues to be used in research to validate NextGen sequencing results.(Matthijs et al., 2016) Why Sanger sequencing did not detect this deletion on this initial test is unclear. Possible contributors include large insertion disrupting the Sanger read of the gene, rearrangement was not amplified, relatively small size of deletion (38bp), or the normal result was due to Sanger read of the intact copy, with the variant copy’s read discarded due to perceived technical error given complex rearrangement.

The CMA completed in 2013 was re-analyzed for this case report to look for evidence of this finding via the CytoChip 180k oligo array platform¹. The raw CMA data does show a ~74kb gain of partial BCL11A, however, this was not reported as the pathogenic effect of a potential interruption of this gene was not defined until several years later. In addition, this gain was below established reporting thresholds (typically 400kb). Additionally, the small deletion in exon 4 of MECP2 caused by the insertion of BCL11A material was below the size threshold of detectable deletions by the microarray platform at the time. As CMA testing does not provide information on the physical location of duplicated genetic material within the genome, there was no evidence to connect the observed BCL11A duplication to the patient’s MECP2 gene. Hence, CMA testing was unable to locate the BCL11A duplication within MECP2 and the MECP2 deletion was too small to be called making the CMA test insensitive, at the time, to this rearrangement. While duplications of BCL11A may be pathogenic (Peron et al., 1993) the impact of this partial duplication is unknown. FISH studies of MECP2 were not performed but are unlikely to have yielded helpful results as the majority of the MECP2 gene has a typical copy number and the insertion of partial BCL11A may not be sufficient in size to disrupt hybridization.

The second sequencing of the MECP2 gene through Greenwood Genetics in 2013 was completed via SOLiD sequencing on a NexGen platform. Technology at that time was limited by fairly short reads (75bp)(Kchouk et al., 2017) and decreased sequencing depth, making it challenging to capture small deletions/duplications, especially complex rearrangements. Supplemental figure 1 is the Greenwood Genetics Variant Call Format File for MECP2 at the time of the sequencing. The highlighted portion is the region of deletion in our patient. The number of reads/coverage does decrease in this area, with the smallest amount of reads being 17, which could have contributed to the missed deletion.

However, medium sized insertion/deletions (10-1000bp) are still challenging to recognize on today’s NextGen sequencing platforms.(Abel & Duncavage, 2013; Spencer et al., 2013) Depending on the coverage depth it would still be possible to miss this deletion in MECP2 today. This case underscores the importance of evaluating a test’s competency in reporting deletions/duplications, as well as understanding the changing landscape of technology. Depending on when a genetic test was completed it may be worthwhile repeating sequencing with updated technologies/systems. While NextGen sequencing has paved the way for rapid assimilation of genomes and advances in genetic research, it does have limitations, which need to be considered. Those limitations will likely continue to decrease with improving technology, such as long read technology, which could potentially alleviate some of the limitations discussed above. Therefore, it is vital to consider when genetic testing was performed in an individual and decide if repetition is warranted should the diagnosis remain unclear.