Retraction Controversy

In the September 23, 2005 issue, Cell published a re-traction of a 2004 paper from Nitz et al. on heritable integration of parasite DNA into host genomes. This was a very unusual situation in which the paper was retracted by the journal rather than by the authors of the paper. The decision to retract the paper came after a year of careful consideration and extensive correspondence with both the original authors and independent experts in the field, including one of the original reviewers, in response to concerns initially raised by a third party. In the end, the reviewers strongly and unanimously recommended that the paper be retracted. The authors were presented with the reviewers' comments and were encouraged to author the retraction, but they declined to do so. Based on the findings of the expert evaluations, we felt that the only responsible course of action was to issue an editorial retraction of the paper. Since the publication of the retraction, we have received letters expressing concern about this action. One of the primary issues raised was that the text of the retraction does not provide sufficient information or scientific justification for the decision. In the following comments, I hope to provide greater clarity on both the scientific issues and the editorial evaluation process that prompted the retraction and to address some of the broader issues that have come to light following its publication. The Nitz et al. (2004) paper reported three main findings. The first was that mitochondrial kinetoplast DNA (kDNA) from the parasite T. cruzi is integrated into the host genome in humans, rabbits, and chicken. The second was that in human Chagas patients, the β-globin locus and LINE-1 retrotransposons were frequently targeted insertion sites. And the third was that kDNA incorporated into the chicken germline was inherited in the absence of persistent T. cruzi infection. Two lines of experimental evidence were used to support the claims of integration and heritability. One was a series of Southern blot analyses of digested host genomic DNA labeled with a T. cruzi kDNA minicircle probe, which revealed kDNA-containing DNA fragments unique to infected tissues. The second line of supporting experimental evidence came from 5# RACE analyses of the host DNA flanking the kDNA sequences. The paper also reported on acute T. cruzi infection of rabbit embryo stem cells and endodermal and mesodermal cells from chicken embryos and presented data on pathogenic myocarditis …