Improvement of a Rapid and Highly Sensitive Method for the Diagnosis of the Mitochondrial m.1555A>G Mutation Based on a Single-Stranded Tag Hybridization Chromatographic Printed-Array Strip

Introduction

Among the pathogenic variants in mitochondrial DNA, the m.1555A>G mutation is one of the most common genetic causes of maternally inherited hearing loss (HL) (Usami et al., 2000b; Jacobs et al., 2005; Yano et al., 2014; Usami and Nishio, 2018) and it is also known to increase the susceptibility to aminoglycoside antibiotics and the risk for aminoglycoside-induced HL (Prezant et al., 1993). In general, aminoglycoside-induced HL is bilateral and severe to profound, occurring within several days after the administration of aminoglycoside antibiotics (Usami and Nishio, 2018).

Approximately 3% of outpatients and 10% of cochlear implant patients have a mitochondrial m.1555A>G mutation (Usami et al., 2000b). In addition, 0.1-0.7% of the general population from various ethnic backgrounds, also carry this mutation (Usami and Nishio, 2018; Maeda et al., 2020). The mitochondrial m.1555A>G mutation results in conformational changes in the stem-loop structure of the aminoglycoside binding site of mitochondrial 12S rRNA similar to that of the bacterial ribosome. It also decreases protein synthesis in the mitochondria and causes sensorineural hearing loss (SNHL) (Hobbie et al., 2008; O'Sullivan et al., 2017). We previously found that the patients with the m.1555A>G mutation who were exposed to aminoglycoside showed more severe HL (Usami et al., 2000a; Lu et al., 2009). Currently, we distribute a warning card to family members who have not been exposed to aminoglycoside antibiotics (Usami et al., 1999). However, to prevent aminoglycoside-induced HL, a rapid companion genetic screening system is desired.

To date, various genotyping methods for mitochondrial m.1555A>G mutations, including the polymerase chain reaction with restriction enzyme fragment length polymorphism (PCR-RFLP) method (Usami et al., 1997, 1999), Invader assay (Abe et al., 2007), TaqMan genotyping (Huang et al., 2015), multicolor melting curve analysis (Wang et al., 2017), DNA microarray (Lévêque et al., 2007), and next-generation sequencing (Zhong et al., 2013), have been reported. However, most of these methods require special equipment or procedures, such as real-time PCR or next-generation sequencing. In addition, some methods are quite time consuming; for example, PCR-RFLP requires 24 h or more, and a more rapid diagnostic method is desired for companion diagnosis.

Recently, genetic diagnosis of this mutation based on a single-stranded tag hybridization chromatographic printed-array strip (STH-PAS) genotyping method has been reported. This method is easy and rapid, allowing genetic testing using only a PCR system with a step-up cycling function (Tian et al., 2014; Kumondai et al., 2018).

In this study, we report an improved protocol for this STH-PAS-based method for m.1555A>G mutations that enables it to be performed on a standard PCR system and show the accuracy of this method based on a large number of DNA samples collected in our laboratory.

Subjects and Methods

Primer design and genotyping of the m.1555A>G mutation

The primer sets for PCR amplification used in this study are listed in Table 1. Each reverse primer was labeled with a tag-spacer sequence at the 5′ end for detection with PAS, whereas the forward primer was labeled with biotin at the 5′ end (Fig. 1).

OPEN IN VIEWER

FIG. 1.

Schematic diagram of the STH-PAS-based genetic analysis and results of polymerase chain reaction amplicon product signals detected by STH-PAS. Wt: DNA sample without mutation (A residue), Vr: DNA sample with a mitochondrial m.1555A>G mutation (G residue). STH-PAS, single-stranded tag hybridization chromatographic printed-array strip.

OPEN IN VIEWER

Table 1.

Polymerase Chain Reaction Primers Used in This Study

Position	Primer name	Primer sequences (5′-3′)
mtDNA m.1555A>G (rs267606617)	F-1_mt1555A_Wt	[Tag 1]-spacer-TTACCATGTTACGACTTGTC
	F-4_mt1555G_Vr	[Tag 4]-spacer-TACCATGTTACGACTTGCC
	Bio_mt_1265F	[Biotin]-CATCTTCAGCAAACCCTGAT

PCR amplification was performed with 6.15 μL distilled water, 1.0 μL of 10× PCR buffer (with MgCl₂), 0.8 μL of dNTP Mix (2.5 mM each), 1.5 μL of primer mixture (10 pM each of Bio_mt_1265F, F-1_mt1555A_Wt, and F-4_mt1555G_Var), 0.05 μL of Takara Taq (TaKaRa Bio, Shiga, Japan), and 0.5 μL of genomic DNA samples (concentrations of the DNA tested here were 250 ng/μL). The PCR temperature profiles involved an initial denaturation step at 94°C for 2 min; 30 cycles of denaturation at 94°C for 5 s, annealing at 60°C for 20 s, extension at 72°C for 10 s; and a final extension at 72°C for 2 min. The PCR was performed using a TaKaRa PCR Thermal Cycler Dice TP650 (TaKaRa Bio). STH-PAS diagnosis was performed using 20 μL of reaction solution containing 8.0 μL of distilled water, 10 μL of separation buffer (Tohoku Bio-Array, Sendai, Japan), 1.0 μL of streptavidin-coated blue latex suspension (Tohoku Bio-Array), and 1.0 μL of PCR product. The C-PAS4 membrane stick (Tohoku Bio-Array) was dipped into the mixture for 5 min at room temperature. The detailed mechanism for visualization with STH-PAS is reported in Kumondai et al. (2018). The diagnosis of each sample was based on the STH-PAS results (Fig. 1). The blue signals in different positions showed different genotypes in the mitochondrial 1555 residue. Some samples showed double signals on the PAS (in which blue bands were observed for both the wild type and variant), reflecting the heteroplasmy of this mitochondrial DNA position.

Results

Pilot studies showed that this STH-PAS-based method works well even under different concentrations of template DNA, different annealing temperatures, and using different PCR thermal cyclers (Supplementary Fig. S1). The results of the comparison between the STH-PAS-based method and Invader assay for the mitochondrial m.1555A>G mutation are summarized in Table 2. Most of the results obtained with this STH-PAS-based method were correct. There were two false negative but no false positive cases. One of the two false negative cases carried the m.1555A>G mutation as heteroplasmy in very low frequency. A double signal for both the wild-type and mutated sequence was observed in five samples. All of these five samples carried this mutation and the wild type with various heteroplasmy ratios, as confirmed by Invader assay. Among 177 samples with the mitochondrial m.1555A>G mutation, 6 samples were heteroplasmic. The sensitivity of this method was 0.99 and the specificity was 1.00. The false negative and false positive ratios were 0 and 0.01, respectively.

This STH-PAS-based method provided results within a 2-h turnaround time including the extraction of DNA from blood or oral epithelial cells, the preparation of the master mix, and the PCR amplification and visualization processes.

Discussion

In this study, we developed a rapid, robust, and easy-to-use diagnostic method for the mitochondrial m.1555A>G mutation. Based on the results of our randomized blind comparison, we concluded that this method is sufficiently reliable for clinical diagnosis and is robust enough for use in different clinical settings. Thus, it will be useful as a companion diagnostic tool for aminoglycoside antibiotics use.

Among the 177 positive control samples, 5 showed double signals (a mutation signal and a wild-type signal) on the STH-PAS analysis, and all 5 of these cases carried the mutation at various heteroplasmy ratios, as confirmed by Invader assay. In previously reported cases, the m.1555A>G mutation was identified as homoplasmic in most cases and only a limited number of patients carried this mutation as heteroplasmy (del Castillo et al., 2003; Lu et al., 2009).

In this study, we could not detect the mutation in two samples. One had a heteroplasmic mutation with a lower mutation frequency, whereas the other case had a homoplasmic mutation. A total of 6 of the 177 positive samples carried this mutation as heteroplasmy. The heteroplasmic ratio of the six samples analyzed here were 1.4%, 7.6%, 8.1%, 23.6%, 27.9%, and 43.5%, with one false negative sample carrying this mutation at a 1.4% heteroplasmic ratio, so we considered this false negative was caused by this very low heteroplasmic ratio under the detection threshold for the STH-PAS-based method. A previous report showed that there is variability in the heteroplasmic ratio of this mutation and this variability is proposed as one of the causes for its different clinical characteristics (Rossignol et al., 2003). In general, patients with this mutation at a lower heteroplasmic ratio would be regarded as at lower risk for aminoglycoside-induced HL. However, the detailed threshold is unknown. As the heteroplasmic ratio may differ among tissues, and the DNA samples could not be extracted from the cochlea, it would be important to find patients with a lower heteroplasmic ratio and instruct them to avoid aminoglycoside use. For the other false negative sample with a homoplasmic mutation, we performed direct sequencing analysis but could not identify any mutation in the PCR-amplified region or its peripheral region. Although unknown factors may be involved in this particular case, this method shows sufficient reliability for companion genetic testing.

The mitochondrial m.1555A>G mutation increases the risk for aminoglycoside-induced HL, but aminoglycoside antibiotics are still used in many countries, especially for patients with tuberculosis, methicillin-resistant Staphylococcus aureus infections, or neonatal intensive care unit babies (Van Boeckel et al., 2014; Krzyżaniak et al., 2016). Meanwhile, newly developed aminoglycoside antibiotics have been proposed (Becker and Cooper, 2013). However, careful attention is needed even with the newly developed aminoglycoside antibiotics, and it has been reported that these drugs are also responsible for HL (Usami et al., 1998). Thus, a rapid, simple, and reliable companion diagnostic method before the use of aminoglycoside antibiotics is important.

This STH-PAS-based method for identifying the mitochondrial m.1555A>G mutation has several advantages in comparison with other methods. This method only requires a commonly used standard PCR thermal cycler (without step-up function). In addition, this method requires little technical experience and provides results within 80 min (including the time for preparation of the master mix and the visualization processes). In this study, although there was no amplicon product contamination, it would be possible to improve our STH-PAS-based diagnostics by employing dUTP and uracil DNA glycosylase for the PCR step to prevent such contamination.

In addition, as this method only requires a PCR thermal cycler, it has been reported that the cost of this method is ∼$7 per sample (Kumondai et al., 2018). We believe this method based on STH-PAS will be beneficial as a tool for companion diagnosis in many countries as this mutation was identified in many ethnic populations (e.g., the Japanese, Mongolian, Chinese, Turkish, Tunisian, Greek, Polish, Spanish, Italian, and American populations) (Usami and Nishio, 2018).