Diagnostic testing for IDH1 and IDH2 variants in acute myeloid leukemia an algorithmic approach using high-resolution melting curve analysis

Abstract

Isocitrate dehydrogenase 1 (IDH1) and IDH2 mutations and polymorphism are reported in 5% to 15% of acute myeloid leukemia (AML) cases, with G105 and R132 of IDH1 and R140 and R172 of IDH2 known to be clinically significant. Current Sanger sequencing assays to detect IDH mutations are labor intensive and not cost effective for clinical testing of low-frequency mutations. Therefore, we developed clinical assays using high-resolution melting (HRM) analysis to screen for all four variants listed above, followed by Sanger sequencing confirmation. The sensitivities of the assays were 7.3% and 7.9% for the detection of IDH2 and IDH1 variants, respectively, against the background of wild-type transcripts. Comparison of HRM to Sanger sequencing on 146 AML bone marrow samples for validation showed near-perfect concordance for all positive and negative results for IDH1 (98%) and IDH2 (94%). Postvalidation clinical implementation of upfront HRM screening (N = 106), using a more conservative algorithm to avoid false-negative results, reduced the number of Sanger sequencing tests by 73% (IDH1) and 78% (IDH2). Of the variant calls made by HRM in postvalidation clinical samples, Sanger confirmed the presence of a variant in 62% (IDH1) and 44% (IDH2) of the samples. In conclusion, our HRM assays are rapid, convenient, and versatile assays for screening and confirmation of alterations in IDH1 and IDH2.

Figure 1

PCR strategy to cover important exon 4 mutations in IDH1 and IDH2. Sequences (bold letters) of (A) IDH1 exon 4 (NCBI Reference Sequence: NC_000002.11) and (B) IDH2 Exon 4 (NCBI Reference Sequence: NC_000015.9) are shown with flanking intronic sequences. The alignment of forward (gray shaded, black fonts) and reverse primers (black shaded, white fonts) show that partial IDH1 exon 4 and entire IDH2 exon 4 are amplified. The boxes mark arginine residues at the active sites for IDH1R100 (CGA), IDH1R109 (AGA), IDH1R132 (CGT), IDH2R140 (CGG), IDH2R149 (CGG), and IDH2R172 (AGG). The location of a known SNP at IDH1G105 is underlined.

Figure 2

High-resolution melting curve analysis for IDH1G105 in duplicate samples. The differential melting properties of representative wild-type and polymorphic IDH1G105 amplicons are shown using (A) normalized and temperature-shifted melting curves, (B) derivative plots, and (C) normalized and temperature-shifted difference plots. WT, wild type.

Figure 3

Detection of IDH1 mutations by HRM analysis. Difference plot analysis representing duplicate samples shows wild-type IDH1 sequence as a flat line. The samples with R132 mutation, G105 polymorphism, and coexisting R132 mutation and polymorphism in G105 show characteristic patterns as marked.

Figure 4

Detection of IDH2 mutations by HRM analysis. A: The derivative plot analysis of IDH2 amplicons shows bimodal peaks. B: Mutations in IDH2R140 and (C) IDH2R172 are detected on difference plot analysis of the left peak and the right peak, respectively.

Figure 5

Confirmation of variant HRM sequences by Sanger sequencing. Representative examples of mutations are shown in IDH1G105 (A–D), IDH2R132 (E–H), IDH2R140 (I–L), and IDH2R172 (M–P). The HRM analysis shows a characteristic shift in the melting pattern of amplicons by derivative plots (A, E, I, M, Q; y axis depicts −(d/dt) fluorescence) and difference plots (B, F, J, N, R; y axis depicts relative signal difference). The Sanger sequencing performed in parallel confirmed the presence of mutations (C, G, K, O, S). A representative wild-type sequence from a different patient sample (D, H, L, P, T) is shown for comparison. A representative patient sample showing successful Sanger sequencing confirmation using post-HRM DNA is shown in the bottom row (Q–T).

Figure 6

Sensitivity of detecting IDH1 and IDH2 mutations by HRM. Normalized and temperature-shifted difference plots for (A) IDH1G105 and (B) IDH2R172 were generated by diluting mutated patient sample DNA into wild-type patient sample DNA. A wild-type patient sample DNA was used as a negative control. The sensitivity adjusted for blast count is 7.9% polymorphic alleles for IDH1G105 and 7.3% mutant alleles for IDH2R172.