Allelic Dropout Can Cause False-Positive Results for Prader-Willi and Angelman Syndrome Testing

Abstract

The diagnosis of many genetic disorders relies on a combination of clinical suspicion and confirmatory genetic testing. Our laboratory uses a standard methylation-sensitive PCR (MSP) to target the differentially methylated SNRPN gene to test for Prader-Willi syndrome (PWS) and Angelman syndrome. One patient, a 27-month-old female, who lacked the classical clinical features of PWS, but had a molecular diagnosis of PWS by MSP by another laboratory, had repeat testing in our laboratory. Testing by MSP in our laboratory also identified an apparent loss of the unmethylated paternal allele, consistent with a diagnosis of PWS. Confirmatory testing using Southern blot analysis with a methylation-sensitive restriction enzyme showed a normal pattern of methylation, detecting both the methylated maternal and unmethylated paternal alleles. To investigate these discrepant results, we amplified and sequenced the SNRPN locus in this patient and identified a single nucleotide change within the binding site for the unmethylated DNA-specific primer. These results indicate this nucleotide change led to allelic dropout in the MSP analysis, yielding the false-positive result. Subsequently, MSP analysis using an alternate primer set that was developed by our laboratory detected both methylated and unmethylated alleles. These findings illustrate that allelic dropout due to the presence of rare polymorphisms can cause false-positive results in commonly used MSP assays and lead to molecular misdiagnosis.

Prader-Willi syndrome (PWS) is characterized by severe hypotonia and feeding difficulty in early infancy, and cognitive impairment, hypogonadism, and hyperphagia leading to obesity in late infancy. PWS is caused by loss of one or more paternally inherited transcripts at the imprinted 15q11-q13 locus. In approximately 70% of patients, this loss is due to a large interstitial deletion of ∼5 megabases on the paternally derived chromosome.¹ An additional 25% of patients have maternal uniparental disomy for chromosome 15, and 1% to 2% of cases have much smaller deletions that are thought to encompass a bipartite imprinting center within the Prader-Willi critical region (PWCR).^2,3 Several genes in the PWCR locus are subject to imprinting, including the SNURF-SNRPN gene, which encodes a single transcript used to transcribe the SNURF and SNRPN polypeptides.^4,5 Normally, SNURF and SNRPN are transcribed from the paternal unmethylated allele only and are silenced in the methylated maternal allele.¹ Because loss of gene expression from the paternally inherited chromosome leads to PWS, SNURF-SNRPN has been a candidate gene for this disease.⁶ Recently, Sahoo et al⁷ showed that loss of expression from adjacent snoRNA genes is implicated in the etiology of PWS, as a microdeletion on the paternal allele that does not encompass the SNURF-SNRPN gene was identified in a patient who met the clinical criteria of the disease.

Angelman syndrome (AS) is also caused by disruption of gene expression in the PWCR. AS is characterized by severe developmental delay, speech impairment, gait abnormalities, and a characteristic happy demeanor.⁸ In contrast to PWS, AS is due to the absence of gene expression from the maternally inherited allele.³ Large interstitial deletions and paternal uniparental disomy account for 80% of AS cases. Mutation of the maternally inherited UBE3A gene, located within the PWCR, is found in 5% to 11% cases of AS, indicating that loss of expression of this single gene is responsible for the phenotype.⁹ The remaining cases of AS have no identifiable defect at 15q11-q13; however, some clinically diagnosed AS patients have been found to carry mutations in the MECP2 gene on the X chromosome.¹⁰

Until the late 1990s, molecular diagnosis of PWS and AS relied on Southern blot analysis of genomic DNA digested with a methylation-sensitive restriction enzyme and a non–methylation-sensitive restriction enzyme that excises the SNRPN promoter region (Figure 1A). In normal individuals, two distinct bands are visualized after transfer and hybridization with an SNRPN probe. The absence of the restriction fragment corresponding to the paternal unmethylated allele is indicative of a diagnosis of PWS; reciprocally, the lack of the maternal methylated allele is consistent with an AS diagnosis.

Figure 1.

A: Schematic map of the PWCR. Physical map of the SNRPN, UBE3A, and snoRNA HBII-85 loci. Also shown is the general layout of the sequence corresponding to the primers and restriction sites for MSP and Southern blot analyses, respectively. B: MSP analysis of SNRPN DNA of patient samples. PCR products of 174 bp and 100 bp are amplified from methylated and unmethylated alleles of the SNRPN DNA locus, respectively. The samples are as follows: lane 1, 100-bp NEB ladder; lanes 2 and 3, patient samples (lane 3 contains the sample with allelic dropout); lane 4, negative control; lane 5, PWS positive control; lane 6, AS positive control; lanes 7 and 9 are empty; lane 8, no template control; and lane 10, untreated genomic DNA control. C: Southern analysis of SNRPN DNA of patient samples. Restriction fragment analysis of patient DNA digested with the methylation-sensitive enzyme KspI and the methylation-insensitive enzyme XbaI. The methylated maternal-specific band is 4.2 kb, and the unmethylated paternal-specific band is 1.1 kb. The samples are as follows: lanes 1 and 8, NEB 1-kb ladder; lanes 2 and 3, patient samples (lane 2 contains the sample with allelic dropout and lane 3 contains a sample from a PWS patient); lane 4, negative control; lane 5, PWS positive control; lane 6, empty; lane 7, AS positive control.

Southern blot analysis is a time-consuming and labor-intensive method that requires the use of radioactive isotopes. As an alternative to Southern blot analysis, Kubota et al¹¹ developed a methylation-sensitive PCR (MSP) assay that targets the SNRPN promoter. In MSP analysis, genomic DNA is treated with sodium bisulfite, which deaminates nonmethylated cytosines to form uracil. Methylated cytosines are resistant to this chemical conversion. Two pairs of primers that hybridize specifically to either methylated or unmethylated SNRPN DNA after sodium bisulfite treatment are used in a multiplex PCR reaction to amplify the two parental alleles.^11,12 MSP is faster, more efficient, and more cost effective than Southern blot analysis. For these reasons, MSP has been adapted by many clinical laboratories, including ours, for PWS/AS testing.

Allelic dropout due to single nucleotide polymorphisms (SNPs) in the PCR primer binding sites is a limitation inherent to all PCR-based assays. Because the PWS/AS MSP assay scores the absence of amplification as a positive result, the presence of a SNP in the primer binding site can lead to failure of amplification and a false-positive result for either PWS or AS, depending on which parental chromosome carries the nucleotide change. Here, we describe a false-positive result for PWS resulting from allelic dropout due to the presence of a SNP on the paternal chromosome. This case reinforces the importance of clinical judgment in the diagnosis of patients. In addition, from the laboratory perspective, this case illustrates how allelic dropout can cause a molecular misdiagnosis and that alternative primer sets hybridizing to different sites within the targeted region should be used to confirm positive test results that rely on an absence of PCR amplification.

Materials and Methods

Methylation-Sensitive PCR

DNA was extracted from whole blood using the Puregene DNA extraction kit (Qiagen, Valencia, CA) according to the manufacturer's recommendation. Heparin-free purple-top blood collection tubes were used to avoid inhibition of DNA methylation-sensitive enzymes. Isolated DNA from the patient, positive and negative control samples, and a water-only control were treated with sodium bisulfite (Cat# Fisher S654-500; Fisher Scientific, Pittsburgh, PA) to convert cytosines, but not 5-methylcytosines, to uracil.^13,14 Briefly, 1 μg of DNA was denatured with 0.2 N sodium hydroxide for 10 minutes at 37°C and then mixed with 15 μL of 10 mmol/L hydroquinone and 250 μL of sodium bisulfite followed by an overnight incubation at 54°C. The treated DNA was cleaned up using the Wizard DNA Clean-up System (Cat# Fisher PRA2361) and desulfonated with 3 N sodium hydroxide, precipitated with 95% ethanol, and washed before elution. The converted DNA was used as a template in a multiplex PCR reaction. Untreated DNA served as an additional control. The primer cocktail used in amplification of the DNA has been previously described by Kubota et al.¹¹ Briefly, the methylated SNRPN locus was amplified with the primers MF1 (5′-TAAATAAGTACGTTTGCGCGGTC-3′) and MR1 (5′-AACCTTACCCGCTCCATCGCG-3′) to generate a 174-bp product. A second set of primers, PF2 (5′-GTAGGTTGGTGTGTATGTTTAGGT-3′) and PR2 (5′-ACATCAAACATCTCCAACAACCA-3′), were used to amplify 100 bp of the unmethylated allele. As a control to ensure the sodium bisulfite conversion was complete, a third primer pair specific for unmodified DNA yielding a 200-bp product was included in the multiplex reaction. The SNRPN locus was amplified with the primer set SNRPNF (5′-GGAGGGAGCTGGGACCCC-3′) and SNRPNR (5′-GAAGCCACCGGCACAGCT-3′) from the untreated DNA specifically. HotStarTaq (Qiagen), 10× PCR buffer without MgCl₂, 25 mmol/L MgCl₂, and 1.25 mmol/L dNTPs standard mix (Roche, Indianapolis, IN) was used for all PCR reactions, and the products were run on a 3% agarose gel. The 1-kb ladder (Promega Corp., Madison, WI) was used for size determination.

Alternate MSP Assay

For the alternate assay, the MSP assay was performed as described above, except the MF1 and MR1, and PF2 and PR2 primer sets were replaced with Alt-MF (5′-TCGATGGTATTTTGTTCGTTCGTATTGGGGCGC-3′), Alt-MR (5′-CCATATCCCTTACCCACTACGATTACCCCG-3′), Alt-PF (5′-TTGATGGTATTTTGTTTGTTTGTATTGGGGTGT-3′), and Alt-PR (5′-ACCACAGACACCCACAATAAAACCTATCACA-3′). The alternate primers amplify a 152-bp fragment from the methylated allele and a 92-bp fragment from the unmethylated allele.

Southern Blot Analysis

Isolated DNA was digested for 4 hours at 37°C with KspI and XbaI. XbaI digestion results in a 4.2-kb fragment flanking the SNRPN promoter region and was cleaved into 1.1-kb and 3.1-kb fragments with KspI (CCGC/GG), which cleaves within the SNRPN promoter. Methylated DNA is resistant to cleavage by KspI. Digested DNA was run on a 1% agarose gel, and Southern blotting was performed using Hybond-N+ membrane (GE Healthcare Bio-Sciences Corp, Piscataway, NJ) and probed with a SNRPN-specific probe acquired from American Type Culture Collection (ATCC, Manassas, VA) (number 95678, 95679; phagemid clone of human small nuclear ribonucleoprotein polypeptide N “SNRPN.” Database accession number: DNA seq. Acc. L32702). All transfers were done with sodium hydroxide (Cat# Fisher S318-5), and all hybridizations were done in 200 mmol/L Na₂HPO₄ and 2% SDS.¹⁵ The methylated maternal SNRPN allele produces a 4.2-kb fragment, and the unmethylated paternal allele produces a 1.1-kb fragment.

Results and Discussion

A 27-month-old female was referred to our clinic with an established diagnosis of PWS by methylation studies from another laboratory. Aside from hypotonia in infancy and developmental delay, this patient did not present with classic clinical features of PWS. She did have mild dysmorphic features, including relative macrocephaly, almond-shaped eyes, midfacial hypoplasia with anteverted nares, and a tented upper lip, as well as subtle brachydactyly with normal flexion creases, and normal nails except for a dysplastic left fifth toenail. Given her lack of a classic PWS phenotype, repeat testing was requested. SNRPN methylation analysis was carried out by MSP as described by Kubota et al.¹¹ The unmethylated paternal allele was not detected (Figure 1B), consistent with the original result and a diagnosis of PWS. Subsequently, Southern blot analysis was also performed, and detected both methylated maternal and unmethylated paternal SNRPN alleles, consistent with a normal result (Figure 1C).

To resolve the discrepant MSP and Southern blot analysis results, we sequenced the SNRPN promoter region targeted by the MSP primers. A single nucleotide change G>A (Chr15: g.22,751,239) was detected that corresponded to the next to the last nucleotide in the binding site for the reverse primer specific for the unmethylated allele amplification (Figure 2). The presence of this nucleotide change likely inhibited primer binding and resulted in the loss of amplification of the paternal unmethlyated allele. Subsequently, we tested relatives of this patient. Both the father and a younger sibling of the patient carry the nucleotide change and tested positive for PWS on MSP (data not shown). These two individuals do not share the phenotype of the patient, and we conclude that the familial nucleotide change is unrelated to the patient's clinical presentation.

Figure 2.

Nucleotide change detected in the primer sequence specific for amplifying unmethylated SNRPN DNA in a patient sample. Electropherogram demonstrating a heterozygous G to A change in the SNRPN DNA of the proband (in bold, marked with the double headed arrow), corresponding to the next-to-last base of the reverse primer sequence (PR2) used in amplifying the unmethylated allele in the MSP.

We reviewed all cases tested by MSP analysis in our laboratory to determine the frequency of this nucleotide change, or other nucleotide changes in the primer binding sites, that may interfere with amplification. Since the implementation of MSP, we have tested 574 samples for PWS/AS. Both methylated maternal and unmethylated paternal alleles were detected in 547 samples, indicating the absence of interfering polymorphisms; the remaining 27 samples were positive for either PWS or AS. Ten were confirmed to carry the recurrent 5-megabase deletion by fluorescence in situ hybridization or array comparative genomic hybridization analysis, indicating the absence of either the paternal or maternal alleles. The remaining 17 samples were retested by Southern blot analysis. We were able to confirm the diagnosis for all but four samples, which lacked sufficient DNA for Southern blot analysis.

To circumvent this problem in the future and verify that a positive result is truly due to the absence of SNRPN methylation and not due to allelic dropout, an alternate MSP primer set, which targets different CpG dinucleotides within the SNRPN differentially methylated region, was designed (Figure 3A). In the patient with allelic dropout, we detected both methylated and unmethylated alleles, consistent with the Southern blot analysis results (Figure 3B). In our laboratory, this alternate primer set is now used in conjunction with the original Kubota et al primers to independently assess methylation for all patient samples submitted for PWS/AS testing.

Figure 3.

A:SNRPN DNA sequence with primers used for MSP analysis in testing for PWS and AS. The sequence of the SNRPN promoter that displays differential methylation is shown from position 22751100 to 22751646. The top line represents the genomic reference sequence. The second line represents the maternal methylated sequence, and the third line the paternal unmethylated sequence after sodium bisulfite treatment. The CpG dinucleotides are shown in red. Sequences highlighted in gray correspond to primers that were designed originally.¹¹ Sequences highlighted in yellow correspond to the alternative primers designed during this study. Sequences highlighted in magenta correspond to primers that amplify unmodified DNA. The G highlighted in green indicates the location of the SNP causing allelic dropout. B: MSP analysis for PWS and AS using the Kubota et al primer set and an alternative primer set. Top panel: Sodium bisulfite-treated DNA was amplified with original primer sets designed by Kubota et al.¹¹ The bands between the 100-bp band and the 174-bp band are observed frequently with this assay, and we attribute these bands to nonspecific amplification of other regions in the sodium bisulfite-treated genomic DNA. Bottom panel: The same sodium bisulfite-treated DNAs were amplified with the alternative primer set. Lane 1, 100-bp ladder; lanes 2 to 4, patient samples (lane 4 contains the sample with allelic dropout); lane 5, negative control; lane 6, PWS positive control; lane 7, AS positive control; lane 8, no template control; lane 9, untreated genomic DNA.

Methylation studies, often via MSP analysis, are used as an initial test when PWS or AS is suspected. A search of GeneTests revealed that there are 62 laboratories offering methylation analysis for PWS diagnosis, with the majority of laboratories not specifying on their websites the testing methodology used for methylation analysis. However, 17 laboratories do state that MSP is used to test for PWS and AS. Notably, any test that relies on a single primer set to detect the presence of methylated and unmethylated alleles is in danger of generating false-positive results.

MSP is a reliable, simple, and efficient method for the initial evaluation of patients suspected of having PWS or AS. Nevertheless, as is true for any PCR-based assay, a limitation of MSP is that the presence of a nucleotide change can cause allelic dropout, yielding a false-positive result. Primer design always runs the risk of overlapping on a rare, unknown polymorphism. This is of particular significance when an abnormal diagnostic test result is due to a failure of amplification, as in the case of PWS and AS. Similarly, we previously reported a Duchenne muscular dystrophy case where allele drop-out in an multiplex ligation-dependent probe amplification assay resulted in a false-positive result for a single DMD exon deletion.¹⁶ To prevent this problem, redundant or complimentary assays can be used in clinical testing. For example, dual testing is common practice in fragile X testing, where samples are tested by both methylation analysis (MSP or Southern blot analysis) and trinucleotide repeat length analysis (by capillary electrophoresis). Having complimentary assays, as in fragile X syndrome testing, or using alternative primer sets in DNA sequencing assays, or as reported here, with alternate MSP primer sets, can significantly decrease the incidence of such false positives.