To the editor

Initial studies of gene correction using chimeric RNA/DNA oligonucleotides (RDOs), also termed chimeraplasts, reported exciting rates of correction of point mutations (up to 40% in one in vivo model1). However, the technique has not yet fulfilled its promise in the fields of transgenics and gene therapy, and some groups have reported persistent failures2,3. In this context, high hopes for the outstanding potential of the technique were raised again when a group of independent researchers recently reported success4,5. We believe, however, it is wise not only to share negative results, as Van der Steege et al.3 highlighted in the April issue, but also to propose criteria for further experiments intending to validate RDO technology.

We have determined four criteria that are necessary to address the criticisms6,7 and cautionary remarks8 expressed in early reports dealing with RDOs. Briefly these criteria are as follows:

  • First, the mutational activity of an RDO must be established for gene conversion from a wild type to a rare mutant genotype in order to obviate spontaneous reversion events.

  • Second, the specific mutation to be introduced should be absent in the cell line used for experiments. Ideally, cell lines harboring the mutated genotype should not be present in the laboratory, to avoid cross-contamination artifacts.

  • Third, mutated clones must be grown and studied individually, as pooled-cell extracts containing large molar amounts of reagent oligonucleotides are poorly suited to estimate rates of chromosomal gene conversion, as well as to confirm gene conversion by sequencing and to avoid other causes of PCR-based artifacts.

  • Fourth, gene conversion rates should be confirmed at the protein level in pools of cells at different time points after the RDO treatment, in order to check for the artifacts in selection of very rare mutation events through multiple cycles of cell culture.

Designing our experiments with these criteria in mind, we have targeted the hypoxanthine phosphoribosyltransferase HPRT gene. HPRT+ and HPRT− cells can be readily selected in HAT (hypoxanthine + aminopterine + thymidine) medium and in 6-thioguanine (6-TG) medium, respectively. The gene is expressed ubiquitously and is located on the X chromosome so that only one allele is to be mutated in male (XY) cells. We aimed at introducing a mutation previously described in a human patient with HPRT deficiency, G310→C at the level of the HPRT complementary DNA (HPRTYale; ref. 9). This mutation leaves virtually no residual enzymatic activity in functional assays; it produces a messenger RNA (mRNA) in normal amounts yielding a protein recognized by a monoclonal and polyclonal antibodies10, and it generates a restriction site for enzyme HhaI, which eases detection at the genomic level.

A chimeric RDO was designed according to published data11 to produce the HPRTYale genotype. It was introduced in human male, diploid cell lines HCT-15 and HBE4-E6/E7 (ATCC numbers CCL-225 and CRL-2078, respectively) either by transfection with Fugene 6 (Roche, Brussels) or polyethylenimine (Fluka, Brussels), and by direct microinjection. HPRT− mutants were selected in 6-TG medium and resistant clones were isolated. Experimental procedures were as reported12,13. Controls included cells treated with various molar concentrations of a wild-type RDO, and cells exposed to the transfectant reagents alone or microinjected with buffer only.

In all experiments, we failed to observe any significant difference in the number of 6-TG-resistant clones between controls and RDO-treated cells. Moreover, resistant clones did not result from specific mutagenesis but merely from the background noise of the selection process, as the G310→C mutation was not observed either by PCR-RFLP analysis or by direct sequencing of genomic DNA from these clones. Thus, in our hands RDOs were not able to produce a HPRT gene conversion rate above background level.

Experiments meeting the above requirements are mandatory to validate the outstanding rates of gene conversion reported with RDOs. In reviewing the 20 original studies published on chromosomal gene conversion using RDOs, we found none fulfilling all four of our criteria. In view of potential artifacts and the lack of reproducibility of published reports3, we consider that conversion mediated by chimeric RDOs still awaits validation. We hope that confrontation of negative as well as positive results will help solving the problem of reproducibility of this potentially promising methodology.