Addition of H19 'loss of methylation testing' for Beckwith-Wiedemann syndrome (BWS) increases the diagnostic yield

Abstract

Beckwith-Wiedemann syndrome (BWS) is a clinical diagnosis; however, molecular confirmation via abnormal methylation of DMR2(LIT1) and/or DMR1(H19) has clinical utility due to epigenotype-tumor association. Despite the strong link between H19 hypermethylation and tumor risk, several diagnostic laboratories only test for hypomethylation of LIT1. We assessed the added diagnostic value of combined LIT1 and H19 testing in a large series of referred samples from 1298 patients, including 53 well-characterized patients from the St. Louis Children's Hospital BWS-Registry (validation samples) and 1245 consecutive nationwide referrals (practice samples). Methylation-sensitive enzymatic digestion with Southern hybridization assessed loss of normal imprinting. In the validation group, abnormal LIT1 hypomethylation was detected in 60% (32/52) of patients but LIT1/H19-combined testing was abnormal in 68% (36/53); sensitivity in the practice setting demonstrated 27% (342/1245) abnormal LIT1 and 32% (404/1245) abnormal LIT1/H19-combined. In addition, H19 methylation was abnormal in 7% of LIT1-normal patients. We observed absence of uniparental disomy (UPD) in 27% of combined LIT1/H19-abnormal samples, diagnostic of multilocus methylation abnormalities; in contrast to studies implicating that combined LIT1/H19 abnormalities are diagnostic of UPD. The overall low detection rate, even in validated patient samples and despite characterization of both loci and UPD status, emphasizes the importance of clinical diagnosis in BWS.

Figure 1

Normal and abnormal methylation in Beckwith-Wiedemann syndrome. A: Location of markers used for UPD-analysis (filled circles, practice group; open circles, validation group). B: Chromosomal region 11p15.5 involved in BWS-associated genomic imprinting defects. With the exception of IGF2 and LIT1, all imprinted genes are expressed (open boxes) from the maternal allele (arrows in transcription direction); silenced genes, black boxes. C: The region can be divided into a centromeric and telomeric domain. Mitsuya et al described the existence of a long QT intronic transcript (LIT1) within KCNQ1OT1 that was transcribed in antisense orientation and is referred to as KCNQ1 overlapping transcript 1 (KCNQ1OT1). The assay targets a NotI site in the CpG island within intron 10 of the KCNQ1 gene and at the 5′ end of KCNQ1OT1 known as DMR2 (synonyms are IC2, ICR2, BWSIC2 and KvDMR1). H19 is also known as BWS, and the assay targets a SmaI site 5 kb upstream of the H19 promoter between exon 3 and 4 of IGF2 known as DMR1 (synonyms are: IC1, ICR1, BWSIC1, H19DMR, and CTCF binding region). Because DMR2 can refer to a centromeric or telomeric methylation center, in the diagnostic setting LIT1/H19 are the preferred terms. Other differentially methylated regions (DMR) upstream of H19, not assayed here, are depicted (DMR0 located 5′ of the main IGF2 promoter, and DMR2 located between intron 7 and 9 of IGF2¹⁹); composed after previous studies.D: Validation of cloned probe DNA fragments in a known BWS-UPD case. There are no methylated LIT1 and no unmethylated H19 bands; methylation index (MI) = top band/sum approximates 0 and 1, respectively. NL, nonaffected individual; BWS, clinical diagnosis of BWS.

Figure 2

Threshold determination in BWS via methylation hybridization. A: Examples of autoradiographs from the validation group for the two loci. Absence of maternal methylation pattern is more obvious and complete in LIT1 when compared with the mostly incomplete hypermethylation of H19. Methylation index (MI) = top band/sum (see Materials and Methods) provided below. B: Based on the methylation indices for the validation group, diagnostic thresholds for the diagnosis of BWS were determined as <0.38 (LIT1) and >0.65 (H19), corresponding to mean control ± 2 σ (details see methods; P values provided in Table 3). NL, nonaffected individual; BWS, clinical diagnosis of BWS.

Figure 3

Addition of H19 increases diagnostic yield. A: In the validation group, the addition of H19-testing allows detection of four additional patients in the LIT1-normal group, which corresponds to ∼8% increased detection. B: In the practice group, the addition of H19 allows identification of 62 additional patients, missed by LIT1 testing alone. This corresponds to an overall ∼5% increased detection rate. C: Southern blot examples from the practice group illustrate the four possible band combinations and are shown as LIT1 and H19 combinations (although run as separate tests). The diagnostic interpretation (Interpr.) is provided under the methylation index (MI) = top band/sum (see Materials and Methods). +, abnormal; −, normal methylation.

Figure 4

Diagnostic algorithm and molecular diagnostic approach in BWS. A: The diagnostic algorithm follows clinical (phenotypic) features and key examples are provided to illustrate the following interrelation: with decreasing clinical suspicion of BWS, the rate of molecular detection decreases; however, clinical variability or specific settings may trigger testing and confirm abnormal methylation in ‘mild phenotypes’ or hemihyperplasia (for comprehensive coverage of clinical features see Refs.^{78910597172737475}). As a visual estimate, the detection rates in the validation and practice group are provided (gray background). B: Epigenetic testing and assessment of uniparental disomy (UPD) should be combined, and the required samples are listed. Finally, the interpretation of the test result and institution of tumor surveillance procedures requires clinical correlation and additional diagnostic testing (eg, cytogenetics or CDKN1C sequencing).