Authors' Reply

We thank Badve and Narikrishna for their interest in our recent publication entitled ‘Proximity ligation assays for isoform-specific Akt activation in breast cancer identify activated Akt1 as a driver of progression’ [1]. They raise a number of questions relating to the assay method and the conclusions we have drawn. They question the use of threonine 308 as a marker of AKT activation, since ‘Ser 473 is required for full activation’. The activation of Akt involves the phosphorylation of two residues: threonine 308 (Thr308) in the activation loop of the kinase by the protein kinase PDK1, and serine 473 (Ser473) in the hydrophobic motif by the mTORC2 complex; both are required for activation of AKT isoforms [2,3]. One study has demonstrated the phosphorylation of Thr308 to be a more reliable biomarker for the protein kinase activity of Akt in tumour samples than Ser473 [4]. In an ideal experiment, we would have examined the role of both Ser473 and Thr308; however, this was not possible at the time. However, we believe Thr308 is sufficient to reflect activation of AKT isoforms in this case. Contradictory observations have been found in different studies measuring the phosphorylation of Ser473 as a surrogate for AKT activation. Ser473 has been correlated with poor prognosis in breast and ovarian cancer [5,6], whereas other studies have demonstrated this not to be the case [7,8]. We explored the impact of activated AKT1 and AKT2 to study whether isoform-specific activation mediated different effects and, in part, might explain some observed inconsistencies when isoforms are not analysed for this pivotal kinase. We concluded, as noted by Badve and Narikrishna, that AKT1 but not AKT2 phosphorylation was associated with poor outcome in ER-positive breast cancers. Indeed we noted a non-significant trend for AKT2 phosphorylation to reverse the negative prognostic impact of AKT1. Badve and Narikrishna suggest that these results are inconsistent with other publications, and reflect that this is a result of the use of Thr308 as a marker rather than Ser473. We disagree, as outlined above, because Thr308 has been shown to be a reliable biomarker of AKT activity. With regard to the specific points raised by Badve and Narikrishna, which they claim are incompatible with our findings: