In this study, WES was performed for 66 Chinese patients with paediatric-onset MDs. Despite the lack of a local diagnostic facility for a respiratory chain enzyme analysis, an overall diagnostic yield of 35% (23/66) was achieved, which was comparable to that of similar studies in recent years.(7-11)Among the cases identified by WES, 48% (11/23) had mutations in MD genes, while the remaining (12/23) had mutations in non-MD genes.
COQ4 mutation was the most important cause of MDs caused by nDNA mutations in the Southern Chinese patients
Eleven patients (17%) were found to have mutations in MD-related genes, with COQ4 having the highest mutation rate (4/66, 6%). The COQ4 missense variants were classified as VUSs at the initial stage of this study. The pathogenicity was supported by the segregation study and functional analysis of the skin fibroblasts, which showed complex II+III deficiency and low CoQ concentration. COQ4 is associated with primary coenzyme Q10 deficiency type 7.(14) Subsequent to the first Chinese study published in 2019 by Lu et al.(15), which described a pair of siblings with Leigh syndrome caused by homozygous COQ4 c.370G>A, p.(Gly124Ser), our team published the largest COQ4 cases series with 11 Chinese patients from nine families (including patients 1 to 4 in this study) in collaboration with a Taiwanese group.(16) Our collaborative study proved that CoQ10 deficiency could occur in neonates, infants, or children with variable phenotypes, which altered the original notion that CoQ10 deficiency shows only neonatal onset.(14, 17) Furthermore, we found that 10 of the 11 patients carried the c.370G>A, p.(Gly124Ser) mutation. Using a single nucleotide polymorphism (SNP) array, we have previously shown that this mutation is a Chinese-specific founder mutation. In the same year of 2019, Ling et al.(18) also described three unrelated Chinese patients with this founder mutation. Currently, 16 Chinese COQ4 cases have been reported, with 15 patients in 12 non- consanguineous Chinese families carrying the c.370G>A Chinese founder mutation.
Performing multiple studies with patients from the same ethnic background can aid in effectively identifying recurrent founder mutations. In a study by Piekutowska-Abramczuk et al.(19), the c.845_846delCT variant was found in 77.6% of the SURF1 alleles in the cohort of Polish patients; however, it was found in a much lower percentage (9%) in the non-Polish population. In 2013, Pronicka et al.(20) studied a cohort with a similar ethnic background; a homozygous c.1541G>A variant in SCO2 (a known MD gene) was identified in an extremely high proportion of the cohort (35/36). In 2016, the detection of recurrent rare pathogenic variants of FBXL4, ACAD9, and CLPB further extended the scope of the suspected Polish- specific MD mutations. The ethnic specificity (Polish) observed in the MD-causing genes(8) parallels the situation of this study in which the COQ4 variant is strongly enriched in cohorts of Chinese patients. A recent study on IEMs have also identified other founder mutations, e.g. GCDH and SLC25A20, in the Southern Chinese population.(21) It is postulated that, when more cases are identified in the Chinese population, the frequency of the c.370G>A, p.(Gly124Ser) variant or other MD-causing mutations will also subsequently increase. Therefore, we suggest that more studies should be performed in Chinese populations to identify any ethnic-specific variants in MD genes.
Benefits of using WES for first-tier genetic testing
According to the 2015 consensus statement (6), patients should undergo WES only if positive results are not obtained with targeted gene sequencing, gene panel analysis, and mtDNA sequencing. However, in our study, as well as in several previous studies (7-11), WES was chosen as the diagnostic method instead of gene panel–based NGS because of the sensitivity of WES. On comparing genes identified in our first-tier results with genes on MitoCarta, we found that 56% (9/16; one MD and eight non-MD genes) of the genes identified in our study were not included in the MitoCarta gene panel. Furthermore, 44% (7/16) of the genes would have been missed if the mitochondrial disorders panel of Genomics England PanelApp (gene panel curated by UK Genetic Testing Network–Association for Clinical Genome Science)(22) would have been used (seven non-MD genes). If a panel-based NGS method had been adopted instead of WES, 52% (12/23) of the patients found to be positive by WES would not have been identified in our study. This statistic is backed by four studies that were conducted after 2015 and used WES (7-10), wherein 24%–49% of the genes identified would have been missed if gene panel–based NGS based on MitoCarta had been used. The false-negative rates of MitoCarta panel–based NGS range from 16% to 45%.
We compared our study with five other studies that were performed during 2014–2018 and used WES for MD diagnosis.(7-11) A total of 92 MD genes were identified in those five studies; however, only 31 (34%) of these genes were reported more than once. Among those 31 genes, only five were also reported in our study indicating that there was minimal overlap (Table 3). More importantly, all studies identified the disease-causing mutations in non-MD genes in patients with suspected MDs. In our study, 12 patients (18%) had non-MD gene mutations. Non-MD gene involvement has also been significant in previous studies (7-11), ranging from 2% to 19%, but only MECP2 was reported in more than one of the five studies published during 2014–2018 (Table 4). Non-MD genes would not have been detected if only gene panel–based NGS, e.g. MitoCarta, had been used. We therefore recommend WES as a first-tier genetic test for patients with MDs because of the high genetic heterogeneity noted in such cases.
Although analysis of the respiratory chain enzyme activity may provide an effective approach for screening out potential patients with MDs, this test is currently unavailable in Hong Kong. Furthermore, the protocol is not standardized, which may lead to discrepancies because of variation in the assay conditions and control values used by different laboratories.(23) The use of WES for first-tier analysis can help avoid unnecessary invasive biopsies. In our study, 20% of such procedures could have been avoided in one MD and eight non-MD cases because molecular diagnosis was directly established through WES. Furthermore, secondary complex IV deficiency was detected in two non-MD patients with ARX and LDHD mutations. These findings indicate that WES is a more comprehensive diagnostic approach than the use of enzymology alone to diagnose MDs. On the other hand, the traditional way of initially investigating a patient suspected of a MD biochemically followed by WES might reveal enzymatic OXPHOS defects in non-MD genes previously linked to well-defined syndromes but not brought into association with mitochondrial failure before. Understanding such pathophysiology with functional studies would hopefully close this knowledge gap. The lack of clear phenotype–genotype relationships and the fact that ongoing studies constantly identify novel mutations make it inefficient to determine what additional gene panel(s) should be used in conjunction with an MD gene panel. This is because non-MD-associated genes could be involved. The use of WES can help fill in existing knowledge gaps and prevent delay in achieving an accurate genetic diagnosis.
With the improvements achieved in bioinformatics analysis, mitochondrial DNA variants can be identified using the off-target read from WES with high recall rate and precision.(5) Overall, the findings indicate that WES should be used for first-tier genetic testing in our local setting. We also propose that WES results should be incorporated in the diagnostic criteria of MDs.
Limitations of our study and future directions
There are certain shortfalls of using WES. Pathogenic variants located in the non-coding region could not be detected. Single/multiple deletions, or depletion of mtDNA, and low heteroplasmy variants could have been missed.(5) With technological improvements, it is expected that whole-genome sequencing (WGS) will also be used in the future, when reductions in cost and increased accuracy of variant interpretation are achieved.(3)