Reduction of Missed Diagnosis of G6PD Deficiency in Heterozygous Females by G6PD/6PGD Ratio Assay Combined with Amplification Refractory Mutation System PCR

Abstract

Objective

Glucose-6-phosphate dehydrogenase (G6PD) deficiency is an X-linked genetic disorder that results in impaired enzyme activity. The G6PD/6PGD ratio assay was routinely used for G6PD deficiency screening in China, but there is an apparent defect of missed diagnosis in heterozygous females. The study aims to explore the means to improve its accuracy.

Methods

A total of 4,161 Chinese females of childbearing age were collected in this retrospective study. All samples were first subjected to G6PD/6PGD ratio assay and then screened by amplification refractory mutation system PCR (ARMS-PCR) for six hotspot mutants in Chinese population (c.1376G>T, c.1388G>A, c.95A>G, c.1024C>T, c.392G>T, and c.871G>A). For the samples with G6PD/6PGD ratio<1.0 and no mutations were found by ARMS-PCR, next-generation sequencing (NGS) was performed. Sanger sequencing was finally used to verify all the variants.

Results

The prevalence of G6PD deficiency in Shenzhen females of childbearing age was 7.31%. The proportion of the six hotspot mutations accounted for 98.03% of all 304 G6PD variants carriers. Taking the ARMS-PCR/NGS results as a reference, the missed diagnosis rate of the G6PD/6PGD ratio assay was 33.88%. Using ARMS-PCR to retest the samples with a G6PD/6PGD ratio between 1.00 and ∼1.10 or 1.00 and ∼1.15 could reduce the missed diagnosis rate from the original 33.88% to 18.09% or 12.05% separately.

Conclusion

ARMS-PCR is an appropriate supplementary method for discovering most carriers missed by the G6PD/6PGD ratio assay.

Introduction

Glucose-6-phosphate dehydrogenase (G6PD) is a housekeeping enzyme expressed in all body cells, and its prominent physiologic role is to protect against oxidative damage [1, 2, 3]. Around 400 million people worldwide were affected by mutations of the G6PD gene. Over 200 genetic variants of the G6PD gene have been identified, and the frequencies of those mutations vary in diverse populations [4, 5]. In general, individuals with G6PD deficiency will remain clinically asymptomatic throughout all or most of their lives. Some individuals may develop a spectrum of manifestations, including neonatal jaundice, acute hemolytic anemia, and chronic nonspherocytic hemolytic anemia, which is generally triggered by the ingestion of drugs, certain types of foods such as fava beans, exposure to certain substances, or when there is accompanying infection or hypoxia. Due to an X-linked inheritance pattern and random X-inactivation, the enzyme activity in heterozygous females could be either normal or G6PD deficient, posing a significant challenge in diagnosing this disease [6, 7, 8].

Many different diagnostic methods have been used to screen for G6PD deficiency. Biochemical assays mainly include G6PD/6PGD ratio assay [9, 10], quantitative G6PD enzyme activity assay [11, 12], methemoglobin reduction test (MHRT) [13], and fluorescent spot test (FST) [14]. Genetic detection methods are mainly denaturing high-performance liquid chromatography [15], amplification refractory mutation system (ARMS-PCR) [16, 17], high-resolution DNA melting assay [18], and next-generation sequencing (NGS) [19]. Because G6PD/6PGD ratio assay has the advantages of convenience, inexpensive, and automation, it has been widely used in the clinic for large-scale screening in China. However, due to genetic mosaics and acute hemolysis, it has a higher missed diagnosis in heterozygous females. This study's main objective was to evaluate the accuracy of the G6PD/6PGD ratio assay and explore the means to reduce its missed diagnosis, taking the results of Sanger sequencing as a reference.

Materials and Methods

Patient Selection

In this retrospective study, a total of 4,161 Chinese females of childbearing age were collected for the G6PD deficiency screening from the hospital between March 2016 and July 2017 in Shenzhen, China. The average age was 34.5 years old and ranged from 20 to 49. All samples were subjected to G6PD/6PGD ratio assay for a definite ratio and then screened by ARMS-PCR for six hotspot mutants in the Chinese population (c.1376G>T, c.1388G>A, c.95A>G, c.1024C>T, c.392G>T, and c.871G>A). Samples were determined as G6PD deficiency with G6PD/6PGD ratio <1.00, and no mutations found by ARMS-PCR were retested by NGS. Sanger sequencing was finally used to verify all the variants. This study was conducted according to the guidelines of the Declaration of Helsinki. Written informed consent was obtained from the samples before blood sampling, and the Ethics Committee of Shenzhen Health Development Research Center approved this study.

Measurement of G6PD/6PGD Ratio

The G6PD/6PGD ratio was measured using a commercial G6PD biochemistry reagent kit (fluorescence ratio method) (Guangzhou KOFA Biotechnology Co., Ltd.). All of the tests were conducted following the manufacturer's instructions. G6PD deficiency was locally defined as G6PD/6PGD <1.00, while the normal value for adults was defined as G6PD/6PGD ≥1.00.

Genotyping of the Six Common G6PD Variants by ARMS-PCR

Genomic DNA was extracted using the QIAamp DNA blood mini kit (Qiagen, Germany), and quantitative estimation was measured by NanoDrop (Thermo Fisher Scientific, USA). We designed six sets of primers for six G6PD variants using Primer Premier 5.0 according to the NCBI data (G6PD:NG_009015) (online suppl. Table S1; for all online suppl. material, see www.karger.com/doi/10.1159/000527806). PCR amplification was carried out in a 10 μL volume containing 0.2 μL each set of primer (10 μM/L), 1 μL DNA template (40 ng/μL), 6.9 μL double distilled water, 0.3 μL SuperTaq enzyme, and 1 μL self-provisioning 10 × PCR master mix, based on the following reaction conditions: predenaturation at 95°C for 2 min, followed by 30 cycles of 94°C for 10 s, 63°C for 15 s, and 70°C for 10 s; a final extension of 10 min at 72°C completed the reaction. The PCR products were electrophoresed on 1% agarose gel and visualized under ultraviolet light, determining whether a variant was based on each lane's size and the number of bands. Sanger sequencing was used to validate each variant with ABI 3130XL genetic analyzer (ThermoFisher Scientific, USA).

Genotyping of G6PD Variants by NGS

Samples with G6PD/6PGD <1.0 but no mutations found by ARMS-PCR were reexamined by NGS. Primers were designed to amplify the full length of the G6PD gene region by long-range PCR (online suppl. Table S2). The PCR products of each sample were barcoded and pooled for sequencing library construction by using ion plus fragment library kit (ThermoFisher Scientific, USA) and following the manufacturer's protocol. NGS was then performed on an Ion Torrent Proton platform. The Ion Torrent machine transformed electronic sequencing signals into raw DNA sequence data and automatically aligned them to the human genome assembly (hg19) using the embedded TMAP tool. Quality control was performed to determine whether the experiment satisfied further demands. Only samples with >200 × mean depth, >98% coverage of the designed region, and >90% uniformity would pass quality control. Subsequently, embedded plugin TVC was used to call all small variants with default parameters (germline, low stringency), generating VCF output. All the variants were annotated with information such as gene location (RefSeq), gene function, population frequency (East Asian group of 1000 genomes, ExAC, gnomAD), impact prediction (dfnsfp33a, InterVar), and disease database (ClinVar, HGMD pro), through the ANNOVAR pipeline. Importantly, in-house scripts were applied to select nonsynonymous variants and divided them into five categories (pathogenic, likely pathogenic, variant of unknown significance, likely benign, and benign) based on the American College of Medical Genetics and Genomics (ACMG) criteria. Finally, only pathogenic and likely pathogenic variations were retained as disease-causing variants and validated by Sanger sequencing.

Statistical Analysis

Statistical analysis was performed using IBM SPSS statistics 22.0. Descriptive statistics results are shown in the form of a percentage (%), mean ± standard deviation, or frequency. Because the data cannot satisfy the parametric hypothesis of Student's ttest or one-way analysis of variance, a nonparametric test was used to evaluate the mean values of G6PD/6PGD ratio among different G6PD genotypes and disease groups. The difference was statistically significant at p < 0.05.

Results

Consistency between the Results of G6PD/6PGD Ratio Assay and ARMS-PCR/NGS

As shown in Table 1, the G6PD/6GPD ratio assay results showed that 212 cases were G6PD deficient and 3,949 cases were normal. The ARMS-PCR or NGS results showed that 304 cases were G6PD mutation carriers and 3,857 cases were normal. The consistent rate of the above methods was 97.26%. Taking ARMS-PCR/NGS results as the gold standard for G6PD deficiency, the G6PD/6PGD ratio assay had 11 cases of misdiagnosis (0.29%) and 103 cases of missed diagnosis (33.88%). The sensitivity, specificity, positive predictive value, and negative predictive value of the G6PD/6PGD ratio assay were 66.12%, 99.71%, 94.80%, and 2.61%, respectively. Additionally, the area under the ROC curve of the assay was 0.913 (95% confidence interval: 0.887∼0.938), indicating that the results of this assay were still reliable (shown in online suppl. Fig. S1). Together, we can calculate the prevalence of G6PD deficiency in Shenzhen females of childbearing age as 7.31%.

Table 1.

Results of G6PD/6PGD ratio assay and ARMS/NGS

G6PD/6PGD ratio assay	ARMS/NGS		n
	G6PD deficiency	normal
G6PD deficiency (<1.00)	201	11	212
Normal (≥1.00)	103	3,846	3,949
Total	304	3,857	4,161

Detailed G6PD Variants Detected by ARMS-PCR/NGS

Of the 304 cases with G6PD mutations, ARMS-PCR found 298 cases and NGS detected 6 cases. The following was a summary of the ARMS-PCR results: 109 cases of heterozygous c.1388G>A, 100 cases of heterozygous c.1376G>T, 31 cases of heterozygous c.95A>G, 20 cases of heterozygous c.871G>A, 12 cases of heterozygous c.1024C>T, 10 cases of heterozygous c.392G>T, and 16 homozygous or compound heterozygous cases with a combination of the above six variants. Thus, we can calculate that the proportion of the six hotspot mutations accounted for 98.03% of all G6PD variants carriers. Figure 1 was the agarose gel electrophoresis image by ARMS-PCR. The NGS detailed results were as follows: 1 case of heterozygous c.835A>T, 2 cases of heterozygous c.1360C>T, 1 case of heterozygous c.517T>C, 1 case of heterozygous c.187G>A, and 1 case of heterozygous c.703C>T. Sanger sequencing results were 100% consistent with ARMS-PCR or NGS, as shown in Figure 2.

Fig. 1.

Agarose gel electrophoresis of the six common variants detected in ARMS-PCR. Lane 1: c.1376G>T heterozygous, lane 2: c.1376G>T homozygous, lane 3: wild-type homozygous, lane 4: negative control, lane 5: c.1388G>A heterozygous, lane 6: c.1388G>A homozygous, lane 7: wild-type homozygous, lane 8: negative control, lane 9: c.95A>G heterozygous, lane 10: c.95A>G homozygous, lane 11: wild-type homozygous, lane 12: negative control, lane 13: c.1024C>T heterozygous, lane 14: c.1024C>T homozygous, lane 15: wild-type homozygous, lane 16: negative control, lane 17: c.392G>T heterozygous, lane 18: c.392G>T homozygous, lane 19: wild-type homozygous, lane 20: negative control, lane 21: c.871G>A heterozygous, lane 22: c.871G>A homozygous, lane 23: wild-type homozygous, and lane 24: negative control.

Fig. 2.

Genotype of six common variants by Sanger sequencing. (a) c.1376G>T heterozygote; (b) c.1376G>T homozygote; (c) c.1388G>A heterozygote; (d) c.1388G>A homozygote; (e) c.95A>G heterozygote; (f) c.95A>G homozygote; (g) c.1024C>T heterozygote; (h) c.1024C>T homozygote; (i) c.392G>T heterozygote; (j) c.392G>T homozygote; (k) c.871G>A heterozygote; and (l) c.871G>A homozygote. All the variants were highlighted in grey blue.

ARMS-PCR Could Find Most Variants within a Certain G6PD/6PGD Ratio

The average G6PD/6PGD ratio of the 11 misdiagnosis cases was significantly higher than that of 201 cases with G6PD deficiency (0.968 ± 0.021vs. 0.664 ± 0.251, p < 0.0001). The minimum, median, and maximum ratios of the 11 cases were 0.93, 0.96, and 0.99, respectively. The average G6PD/6PGD ratio of the 103 missed diagnosis cases was significantly lower than that of 3,846 normal cases (1.225 ± 0.223 vs. 1.373 ± 0.186, p < 0.0001). The minimum, median, and maximum ratio of the 103 cases was 1.00, 1.13, and 1.80, respectively (shown in online suppl. Fig. S2). The above results suggested that the ratios of those misdiagnosed cases concentrated on a particular range.

Among the 3,949 samples judged as normal by the G6PD/6PGD ratio assay, a total of 103 cases were diagnosed as G6PD carriers with mutations found by ARMS-PCR. We analyzed the mutation number detected by ARMS-PCR in different G6PD/6PGD ratio intervals. As shown in Table 2, 48 mutations were discovered from 222 samples with a ratio of 1.00∼1.10. In other words, using ARMS-PCR to reexamine 222 samples in this range could reduce the missed diagnosis rate from the original 33.88% to 18.09%. Similarly, 65 mutations were found from 449 samples with the 1.00∼1.15 interval. Using ARMS-PCR to reexamine 449 samples could reduce the missed diagnosis rate from 33.88% to 12.50%. By this token, ARMS-PCR is a suitable means to find as many carriers as possible within a definite G6PD/6PGD ratio range, which can make up for the G6PD/6PGD ratio assay defects to a large extent in Chinese heterozygous females screening (Excel S1).

Table 2.

The number of mutations detected within different G6PD/6PGD ratio intervals

Mutation	G6PD/6PGD ratio distribution
	n (1.00˜1.10)	n (1.00˜1.15)	n (1.00˜1.20)	n (1.00˜1.25)	n (≥1.00)
Yes	48	65	65	70	103
No	174	384	629	1,017	3,846
Total	222	449	694	1,087	3,949

Discussion

G6PD deficiency is an inherited erythrocyte enzyme deficiency disease and lacks effective treatments so far. Due to a random X-inactivation, mutations in heterozygous females are mosaics, resulting in different G6PD-enzyme activity and symptom severity [20, 21]. So, using the traditional enzyme activity methods to diagnose G6PD deficiency in heterozygous females is defective. In China, the G6PD deficiency prevalence rate was higher in the south of the Yangtze River, ranging from 3.1% to 16.1%, especially in Guangdong and Guangxi [22, 23, 24]. Similar to those studies, the prevalence of G6PD deficiency in south Chinese females was up to 7.31% in this paper, suggesting that early diagnosis is crucial in those areas. Due to the advantages of the technology itself, the G6PD/6PGD ratio assay is widely used clinically for large-scale G6PD deficiency screening. The result is exceptionally accurate in judging hemizygotes males and homozygotes females, but there is a well-known defect that an omission may occur in evaluating heterozygotes females. In this study, the missed diagnosis percent was 33.88%, coherent with the previously reported 35.0% by Lyu et al. [25]. Therefore, finding a supplemental way to mitigate this drawback as much as possible is essential.

ARMS-PCR, also named allele-specific amplification, can screen any known point mutation. For example, it was performed to detect the two most common G6PD mutations, c.1388G>A and c.1376G>T, in Chinese populations in 1999 [26, 27]. The mutation genotype could be determined by the number and length of DNA bands in agarose gel electrophoresis. The technical difficulty and the key to this method's success were the design of high-quality primers, which must amplify both the wild-type and mutation loci specifically and efficiently. In this article, four primers were designed for a specific mutation: one pair of inner primers with opposite directions belonging to different genotypes (wild type and mutation); one pair of outer primers located outside and at different distances from the SNP site. Pairwise combinations of the above primes could amplify up to three DNA products. To prevent nonspecific amplification, we introduced a base that did not match the human DNA sequence at the third position to the 3′ terminal of the inner primers. Among the Chinese G6PD deficiency patients, the top three frequent mutations were c.1376G>T, c.1388G>A, and c.95A>G, followed by c.871G>A, c.392G>T, and c.1024C>T, with slight geographical or ethnic differences [23, 28, 29]. In our previous study of the Shenzhen population, the six dominant mutants, c.1376G>T (35.52%), c.1388G>A (32.91%), c.95A>G (11.11%), c.871G>A (6.48%), c.392G>T (4.04%), and c.1024C>T (3.96%), had a cumulative proportion of 94.02% [30], consistent with the results reported by He et al. [31] and Fan et al. [16]. Therefore, we selected these six mutations as the screening targets for ARMS-PCR. Our big data showed that among the 304 samples with variants, the top six common mutations that ARMS-PCR could detect accounted for 98.03%, which correlated well with those reported in China by other scholars, such as 97.7% [16] or more than 90% of G6PD-deficient alleles [31]. Also, the primers here can be used clinically because the results have been repeated reliably, with excellent specificity and accuracy. By analyzing the variants at different G6PD/6PGD ratio intervals, especially 1.00∼1.10 and 1.00∼1.15, we can see that AMRS-PCR is a suitable complementary method to find out most carriers missed by the G6PD/6PGD ratio assay. Clinicians can choose the samples of which ratio range to retest by AMRS-PCR according to their needs.

Due to the particularity of G6PD deficiency, it is vital to know early whether a parent has the G6PD mutation. Our results demonstrated that the G6PD/6PGD ratio assay could be used for initial G6PD deficiency screening and ARMS-PCR for six mutations in females within a specific ratio. In this way, the missed diagnosis of the former can be minimized, which is beneficial to the carrier and their offspring.

Statement of Ethics

The research was approved by the Ethics Committee of Shenzhen Health Development Research Center (approval No.2016JYZX112). All participants gave written informed consent to participate in this study. All procedures involving human participants were done following the ethical standards of the institutional or national research committee and with the World Medical Association Declaration of Helsinki.

Conflict of Interest Statement

The authors have no relevant financial or nonfinancial interests to disclose.

Funding Sources

This work was funded by the Science, Technology Innovation Committee of Shenzhen Municipality (Grant No. KJYY20180703173402020).

Author Contributions

Shan Duan and Shiguo Chen designed the study and proofread the manuscript. Shiguo Chen, Jian Gao, and Qunyan Wu performed the experiments and wrote the manuscript. Zhaopeng Guo collected the appropriate samples. Kaifeng Zheng, Xi Li, Sheng Lin, and Jindu Su analyzed and interpreted the data. Jilong Yao proposed constructive revision suggestions and proofread the manuscript. All authors reviewed and approved the final manuscript.

Data Availability Statement

All data generated or analyzed during this study are included in this article and its supplementary material files. Further inquiries can be directed to the corresponding author.

Supplementary Material

Supplementary data

Supplementary data

Funding Statement

This work was funded by the Science, Technology Innovation Committee of Shenzhen Municipality (Grant No. KJYY20180703173402020).