Droplet digital PCR as a first-tier molecular diagnostic tool for focal cortical dysplasia type II

Le résumé fourni par la source

We read with interest the recent article by Pirozzi et al.1 demonstrating the utility of droplet digital PCR (ddPCR) for the molecular genetic diagnosis of focal malformations of cortical development (FMCD) including focal cortical dysplasia (FCD), hemimegalencephaly (HMEG) and dysplastic megalencephaly (DMEG). By targeting six recurrent variants in MTOR, PIK3CA and AKT3, Pirozzi et al.1 achieved a diagnostic rate of 24.2% for FCD (8 of 33 individuals) and 81.8% for HMEG/DMEG (9 of 11 individuals). Our research lab has also implemented a similar strategy to rapidly screen for recurrent variants in FCD samples, with favourable outcomes. As a result, ddPCR has become a routine research genetic testing method for FCD samples collected from the Epilepsy Surgery Program at the Royal Children’s Hospital, Melbourne. The purpose of this letter is to provide supporting evidence for the article by Pirozzi et al.,1 and to advocate for the implementation of ddPCR as a first-tier molecular diagnostic tool for FMCD, particularly FCD Type II (FCDII). We present the research findings generated at the Royal Children’s Hospital, Melbourne, and discuss how this approach can be further improved and utilized as a research or clinical grade diagnostic tool. This study was approved by the Royal Children’s Hospital Ethics Committee and all participants provided informed consent. Genomic DNA isolated from fresh frozen brain samples of 49 individuals with FCD were examined. FCD was identified using data from electrophysiology and neuroimaging. Diagnostic histopathology revealed FCDIIa in 29 individuals and FCDIIb in 20. No prior genetic testing was performed. We designed ddPCR probes to target five somatic MTOR variants commonly associated with FCD in the literature, namely variants resulting in the amino acid substitutions p.S2215F, p.S2215Y, p.T1977K, p.L1460P and p.A1459D. We selected the most dysplastic brain sample for analysis for each individual using histopathology data. This is determined by the abundance of abnormal cell types, namely the dysmorphic neurons and balloon cells. The diagnostic rate was 26.5%, with pathogenic variants found in 13 individuals. The variant allele frequency (VAF) ranged from 0.17% to 6.1% (Table 1). None of the variants were detected in blood-derived genomic DNA. Next, we analysed additional brain samples from 10 individuals with pathogenic variants. In 9 of 10 cases, we found that the most dysplastic brain samples, as determined by histopathology, contained the highest VAF (Table 1). Clinical and molecular data of individuals with pathogenic MTOR variants DX = diagnosis; NP = neuropathology; N/A = not applicable. Clinical and molecular data of individuals with pathogenic MTOR variants DX = diagnosis; NP = neuropathology; N/A = not applicable. These results demonstrate that sample prioritization based on histopathology can increase the efficiency of ddPCR testing. We selected one frozen brain sample from each individual for the initial testing, prioritizing those samples with the most dysplastic features. Since pathogenic somatic variants are known to be enriched in the abnormal cells,2,3 we reasoned that the chance of identifying a pathogenic variant can be increased by selecting the sample with the highest density of abnormal cells using visual inspection. Indeed, our follow-up investigation showed that in 9 of 10 cases, the first sample that was selected contained the highest VAF, demonstrating the validity of our strategy. This highlights the importance of having a detailed neuropathological examination defining not only the diagnosis (e.g. FCDIIa or FCDIIb), but also describing the relative density of abnormal cells in each brain sample. In addition to brain samples, Pirozzi et al.1 also analysed peripheral samples including skin, blood, saliva and buccal swab samples, but brain samples provided the highest solve rate. Interestingly, Pirozzi et al.1 found weak or no correlation between the levels of VAF and histopathological abnormalities in three cases, suggesting that this correlation may not be consistently observed in all brain specimens. Nevertheless, our data and the literature2 suggest the most efficient and cost-effective protocol is to screen a single brain sample, preferably the most dysplastic sample available. The diagnostic yield of ddPCR testing is determined by the selection of TaqMan probes that target different variants. While the inclusion of more probes allows more variants to be tested, an excessive number of probes can compromise the rapid nature of the test and diminish its cost-effectiveness. The selection of probes should be guided by the prevalence of each variant in the disease type of interest. Pirozzi et al.1 selected two probes for MTOR (p.S2215F and p.S2215Y), three probes for PIK3CA (p. E542K, p.E545K and p.H1047R) and one probe for AKT3 (p.E17K) based on the literature and the composition of their cohort. This is an effective selection of probes to study DMEG, HMEG and FCD at the same time, resulting in a diagnostic rate of 24.2% for FCD and 81.8% for HMEG/DMEG. We performed a review of the literature to identify variants associated with FCD, HMEG and DMEG (Supplementary Table 1). We determined that targeting five recurrent MTOR variants is the most efficient strategy for FCD. Previous studies showed that MTOR variants account for ∼20–35% of FCD,2,4–6 and the five most common MTOR variants constitute ∼75% of all MTOR variants found in FCD (Supplementary Table 1). The theoretical diagnostic rate of our study is hence ∼15–26%. In agreement with this estimation, we achieved a diagnostic rate of 26.5% for FCD. It is notable that the most prevalent MTOR variant in our study, p.T1977K (found in 8 of 13 positive FCD cases), was not targeted by Pirozzi et al.1 Similarly, we did not target PIK3CA and AKT3 variants, which were identified in four of eight positive FCD cases by Pirozzi et al.1 A combination of both of our approaches may be beneficial for future studies. There are two main limitations associated with ddPCR testing for FMCD. First, this approach is more suitable for FCDII or HMEG/DMEG with FCDII pathology. This is because the genetics of FCDII is better understood, which allows more sophisticated probe selection for ddPCR testing. Furthermore, FCDII has well-defined histopathological characteristics, with compelling evidence that causal variants are enriched in the abnormal cells (dysmorphic neurons and balloon cells).2,3 This helps inform which brain sample is more likely to contain the pathogenic variant at a higher VAF. This level of knowledge is currently not available for other FMCDs such as FCD Type I and FCD Type III. As a result, ddPCR testing for other FMCDs may not provide a desirable diagnostic yield. Indeed, Pirozzi et al.1 did not identify any variants in other MCDs (seven cases) or other cortical lesions (seven cases). Additional research is required to determine if ddPCR testing might have clinical utility for testing other FMCDs in addition to FCD. The second limitation of ddPCR testing is the type of variants that can be targeted. This methodology is suitable for recurrent, gain-of-function variants in proteins that function as positive regulators of the mTOR pathway, such as PIK3CA, AKT3 and MTOR. Other proteins associated with FCD and HMEG, including TSC1, TSC2 and DEPDC5 are negative regulators of the mTOR pathway, typically associated with loss-of-function variants. These loss-of-function variants are rarely recurrent, and as a result, ddPCR testing to screen for variants in these genes is not efficient or cost-effective. A previous study estimated that pathogenic variants in TSC1, TSC2 and DEPDC5 account for 16.1% of FCD/HMEG cases.2 For the foreseeable future, detection of these variants will continue to rely on high depth next-generation sequencing (NGS) methods such as exome sequencing and targeted panel sequencing. Despite these limitations, ddPCR appears to be an effectiv

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.