PCR‐ and Restriction Endonuclease‐Based Detection of IDH1 Mutations

Abstract

Hotspot mutations in codon 132 of the gene encoding isocitrate dehydrogenase 1 (IDH1) have emerged as the most frequent DNA alteration in astrocytomas, oligodendrogliomas and oligoastrocytomas. These mutations have been shown to be of significant diagnostic and prognostic value. So far, assessment of IDH1 mutation relied on DNA sequencing techniques.

We generated a set of primers suitable for endonuclease‐based detection of hotspot mutations in codon 132 of IDH1. This primer set will allow determining these mutations without the need of DNA sequencing. One set of primer sets is designed to detect the presence or absence of IDH1 mutations in codon 132, while the other primer sets individually recognize the R132H, R132C, R132S, R132G and R132L mutations.

Spontaneous mutations in codon 132 of isocitrate dehydrogenase 1 (IDH1) were shown to occur in the overwhelming majority of diffuse astrocytomas, oligodendrogliomas, mixed oligoastrocytomas and secondary glioblastomas, and with a low frequency in primary glioblastomas 1, 3, 7, 10. The largest series so far detected mutations in 311 of 455 (68.4%) astrocytomas, in 226 of 302 (74.8%) oligodendrogliomas and in 179 of 253 (70.8%) oligoastrocytomas (2). Presence or absence of IDH1 mutations is of strong prognostic power in anaplastic gliomas and in glioblastoma because these mutations are associated with a significantly more favorable clinical course 6, 8, 9, 10. Both the high frequency and the clinical importance result in a high demand of molecular analysis of IDH1. Mutations in IDH1 exhibit a very characteristic pattern. The highly conserved isocitrate binding domain contains somatic heterozygous point mutations in codon 132. The most common types and the frequency of these mutations have been described (2). Detection of these mutations so far relied on single‐strand conformation polymorphism (SSCP) (7) or sequencing analysis 1, 5. However, the strictly circumscribed position and the limited number of different point mutations render IDH1 analysis also attractive for polymerase chain reaction (PCR)‐ and restriction endonuclease‐based analysis. Unfortunately, IDH1 does not contain suitable sites for restriction endonuclease‐based analyses of codon 132.

Therefore, we generated mismatched primers to create suitable restriction sites for wild‐type and mutant sequences. All mutations detected in a large series exceeding 1000 gliomas were taken into account (2). This approach, termed derived cleaved amplified polymorphic sequence (dCAPS) analysis, has previously been described (4). Primer design was based on the dCAPS finder 2.0 program (http://helix.wustl.edu/dcaps/dcaps.html). Mutation‐specific restriction sites, suitable endonucleases and mismatch primers are compiled in Table 1. Design of the forward primers for the wild‐type sequence and for the five common mutations depended on the sequence in codon 132. One of two reverse primers was used for all amplifications depending on preference for longer or shorter amplification products. The combination of IDH1 sequence with sequence‐specific forward primer generated a specific restriction site. Amplification with pIDH1f‐R132 and pIDH1r‐132 yielded a fragment with a PvuI site in the wild‐type sequence. Both IDH1 DNA copies from patients with wild‐type sequence were digested, while in patients with IDH1 codon 132 mutations one allele could not be cleaved, resulting in two signals upon separation on 2.5% agarose gels (Figure 1, upper panel). With the other forward primers, restriction sites specific for codon 132 mutations could be generated, resulting in cleavage of the mutated DNA strand (Figure 1, lower five panels). PCR was performed in a volume of 20 µL containing 10 µL HotStar Taq Plus‐Mix (Qiagen, Hilden, Germany), 10 pmol of forward and reverse primer each, 100 ng of DNA and nuclease‐free water. In order to raise the preset MgCl₂ ion concentration to 2.5 mM, 0.4 µL MgCl₂ of a 50 mM stock solution was added. Initial denaturing at 95°C was followed by 38 cycles of 30 s at 95°C, annealing for 30 s at 55°C and extension for 40 s at 72°C with a final extension step at 72°C for 5 minutes. After amplification, 10 µL of the PCR products were digested for 1 h using 2 units of restriction endonuclease and appropriate buffer (Fermentas, St Leon‐Rot, Germany) in a total volume of 30 µL. Incubation was at 37°C for PvuI, BglII and BcuI, and at 55°C for BclI and BseGI. Amplification products measured 260 bp to 263 bp. The sizes of the cleaved fragments are given in Table 1. Separation on 2.5% agarose gels allowed detection of the non‐cleaved fragment and of the larger of the cleaved fragments spanning 226 bp to 240 bp. The most efficient approach for detecting IDH1 hotspot mutations involves an initial PCR generating a PvuI site spanning codon 132 of wild‐type IDH1, but not in mutant alleles. Complete restriction of this PCR product resulting in a single band upon electrophoretic separation confirms wild‐type sequence (Figure 1, upper panel left). Detection of two signals indicates a mutation which consequently can be determined by employing the set of five mutation‐specific mismatched primers (Figure 1, lower panels). Adaptation of the dCAPS method to mutation analyses of IDH1 provides a rapid and an inexpensive approach. This set of primers allows rapid screening of IDH1 mutations or independent confirmation of sequencing data. It is of major advantage that this approach is based on laboratory equipment common in all molecular diagnostic laboratories. Both DNAs from fresh frozen and from paraffin‐embedded tissue were suitable for this protocol; however, DNAs from paraffin‐embedded tissues preferably were amplified with the reverse primer generating the short set of amplification products. For validation, we compared 60 tumors sequenced for IDH1 codon 132, containing 19 wild‐type, 19 R132H, 8 R132G, 7 R132C, 4 R132S and 3 R132L variants, with results obtained by this PCR‐ and restriction endonuclease‐based detection assay. No discrepancies between the tests were observed. We also considered employing amplification refractory mutation system (ARMS) primers directly specific for IDH1 mutations. Such an approach limits the possibility for optimizing primers, because the 3′ region of the primer is defined by the mutated sequence frequently causing the problem of 3′ self‐complementarity. Further, the difference of only a single base between wild‐type and mutant sequences tends to result in unspecific amplification products. In contrast, using our approach allows optimizing the primers because of the possibility to alter sequence specificity for different enzymes and then to select the best suited primer–enzyme combination. Our method has limitations. Most important, DNA needs to be extracted from areas containing a high proportion of tumor cells. This drawback is shared with conventional DNA sequencing because both methods usually rely on comparable amplification procedures. In comparison to DNA sequencing, our method does not detect novel mutations. For example, the R132V mutation described a single time in the literature (1), is not detected. Indeed, we restricted our set of primers to detect only those IDH1 mutations with a frequency exceeding 0.1%.

Table 1.

Primer sequences and restriction endonucleases for analysis of IDH1 mutations in codon 132. Abbreviations: fragm. = restriction fragments; sf = short amplification fragment; lf = long amplification fragment.

Figure 1.

DNA from 12 different patients with confirmed DNA sequence in codon 132 of IDH1 indicated on top was amplified with sequence‐specific mismatched forward primers, indicated on the left, and a universal reverse primer followed by digestion with restriction endonucleases. Two of the patients were wild type (wt), and two patients each carried R132H, R132C, R132G, R132L or R132S mutations. All reactions were performed with the reverse primer generating the longer fragments. Fragment sizes in base pairs are given on the right. The smaller restriction fragments sized 21 bp to 36 bp are not resolved at this condition for electrophoresis.

This approach designed for the detection of IDH1 mutations might also be applied to the detection of alterations in IDH2 (10) by employing suitable primer pairs. However, in our experience, IDH2 mutations are quite rare occurring in only 3.1% of astrocytomas and oligodendroglial tumors (2), and, therefore, we do not include analysis of these alterations in our routine for molecular analyses.