Nature 475, 231–234 (2011); doi:10.1038/nature10167

corrigendum Nature 481, 534 (2012); doi:10.1038/nature10789

In this Letter, we presented findings from experiments using the EJ bladder xenograft cancer model, in which some tumours on some of the animals exceeded the maximum size (15 mm in any dimension) permitted by the Institutional Animal Care and Use Committee (IACUC) at Massachusetts General Hospital (MGH). Therefore, we withdraw the data presented in Supplementary Fig. 9b, as well as in Fig. 2 from the first Corrigendum. Although other tumours were found to exceed the permitted maximum, owing to the degree of tumour size and the circumstances of the experimental procedures, this was considered acceptable as detailed below.

The tumours in eight mice in Fig. 2b, d (same mice in Fig. 1a, b in the Corrigendum) also exceeded the tumour size approved in the IACUC protocol of the principal investigator (PI), and were euthanized 48–72 h after the tumour burden was identified. Although the tumour size limit permitted by the protocol had been reached in the above-mentioned animals, the mice exhibited no other clinical signs associated with humane endpoints due to pain or distress. Given the lack of clinical signs observed and the timing of the clinical presentation, the PI was permitted to maintain the animals for this short time period to support proper tumour collection and fixation. The animals were regularly monitored during this time period. We also noted that we made an error in the first Corrigendum when calculating the volumes for the last time point of Fig. 1b. All measurements are now presented as Supplementary Data.

The in vivo experiments described in Supplementary Fig. 9 were performed between 2007 and 2008. At this time, the IACUC-approved protocol of the PI required daily monitoring of the tumour-bearing animals for clinical signs of distress, and did not require daily measurements of the tumour size. Measurements of tumour sizes were performed at the indicated time points. Consistent with the IACUC-approved guidelines, all animals were euthanized as soon as measurements indicated that the tumours reached the size limit approved in the protocol (15 mm). These mice did not show clinical signs of distress, and thus it only became apparent that the tumours had already exceeded 15 mm when they were measured. Since the completion of this study, the IACUC-approved protocol of the PI was revised, and daily measurements of tumour sizes, in addition to daily observation, are now required. For the xenograft tumour models, measurements were performed on the entire tumour lesion including cases when tumours appeared as aggregates of single nodules (melanoma model in Supplementary Fig. 9e). All measurements are now presented as Supplementary Data.

The MGH IACUC approved the animal models and the general procedures for drug testing used in this work; however, some compounds and cell lines used for the experiments in Fig. 2 and Supplementary Fig. 9 had not received prospective IACUC approval owing to an administrative oversight. All of the materials included in the original Letter and the earlier Corrigendum have since been approved by the IACUC for subsequent experiments. We would like to clarify that the detection of total p53 in Fig. 4c was performed after stripping the membrane following detection of phosphorylated p53 (Ser15 p-p53). In addition, the original paper incorrectly stated that the error bars in Figs 1c, 3a–d, 4a, b, and Supplementary Figs 2b, 3a, 4a, 6, 15c, 18, 20, 21, 26b and 29b were calculated based on three independent experiments. These graphs and the associated error bars represent the results of technical triplicates from one experiment. The error bars in the graphs throughout the main paper and Supplementary Information represent the standard deviation of the mean. The IACUC has reviewed all the data now presented, and confirms that the statements provided in this Corrigendum are accurate. Corrective measures have since been taken to avoid any irregularities happening again. Although the scientific conclusions of the original paper stand, we would like to apologize for the numerous inaccuracies in reporting our data, and for the breach of animal welfare guidelines in some of the original data.