Correction to: Oncogene (2015) 34:3413–3428 , published online 1 September 2014 Following the publication of this article, it was noted a mistake in the Western blot (WB) shown in Figure 3f. The band corresponding to JIMT-1 cells stimulated with HRGβ1 and treated with trastuzumab in the second panel of Figure 3f (WB revealed with the ErbB-2 antibody) was inadvertently duplicated. This mistake occurred when bands in the original WB were spliced to assemble Figure 3f. The authors have provided the correct raw data and a revised Figure 3f which contains a black line indicating that lanes that were non-adjacent in the gel were re-arranged. Therefore, in the revised second panel of Figure 3f it is shown that: (i) protein extracts present in lanes 1 and 2 were adjacent in the gel, (ii) lanes in the middle of the gel were spliced, which is highlighted by a black line (iii) the adjacent lanes 3 to 6 in the gel were assembled along with lanes 1 and 2 to present data in said second panel. The legend in Figure 3f was modified to duly acknowledge the non-adjacent lanes in the original gel. The revised version reads as: Revised Legend to Figure 3. Inhibition of ErbB-2 nuclear localization blocks proliferation in Ttzm-resistant BC. (a) Cells transfected with pcDNA3.1 (left) or hErbB-2ΔNLS (right) vectors were treated as indicated. ErbB-2 (red) and GFP-tagged-hErbB-2ΔNLS (green) were localized as in Figure 2. Solid arrows: hErbB-2ΔNLS-transfected cells; dashed arrows: wild-type cells that did not uptake the hErbB-2ΔNLS mutant. Nuclei were stained with 4’,6-diamidino-2-phenylindole (DAPI) (blue). (b–d) Quantitative analysis of ErbB-2 subcellular localization. Fluorescence intensities of MErbB-2 and NErbB-2 were quantified in 50 cells from each group and are plotted as the ratio of MErbB-2/NErbB-2 (b), the percentage of MErbB-2 (c) or the percentage of NErbB-2 (d) (mean ± s.d.). For # vs *P < 0.001. For b–e vs a, c–e vs b, and d vs c: P < 0.001. For c and d vs e: P < 0.05. (e) Blockade of NErbB-2 localization inhibits in vitro growth. Cells were transfected and treated as indicated. Proliferation and data were analyzed in Figure 2. For b and c vs a, and d vs b: P < 0.001. (f and g) Effect of hErbB-2ΔNLS on ErbB signaling. Cells were transfected and treated as in (e). WB analysis (f) and immunoprecipitation analysis (g) were performed as in Figure 2. The black line in the second panel of (f) was added to show that lanes 1 and 2 were adjacent in the gel, lanes in the middle of the gel were spliced and lanes 3 to 6 were adjacent in the gel. (h) Cell proliferation in intrinsic and acquired Ttzm-resistant cells was evaluated by [3H]thymidine incorporation or cell count. Data are presented as in (e). For b and c vs a, and d vs b: P < 0.001. Experiments in (a–h) were repeated three times, with similar results. (i) hErbB-2ΔNLS blocks in vivo growth. Left panel: Cells were inoculated subcutaneously in mice. When JIMT-1 tumors reached 40 mm3, animals were divided into three groups and injected with Ttzm, IgG, or remained untreated. JIMT-1-hErbB-2ΔNLS-injected mice developed tumors with a latency of 40 days. Each point represents the mean tumor volume ± s.e.m. Right panel: Decrease in tumor mass. (j) Tumor growth. aGrowth rates were calculated as in Figure 1. Volume, percentage of growth inhibition, and delay in tumor growth from the different groups with respect to tumors from JIMT-1 cells were calculated at day 64. # vs *P < 0.001. bWith respect to JIMT-1 cells, P < 0.001 (see Supplementary Fig. S2). (Figure presented.) The authors emphasize that under “Results” they have shown the same results as those presented in Fig 3f for JIMT-1 cells in multiple BC cell lines and tumor models: A. Murine BC C4HD cells, which mimic ErbB-2-positive BC (“Nuclear ErbB-2 drives HRGβ1-induced BC growth”); B. An in vivo BC model induced by HRGβ1 (“Nuclear ErbB-2 drives HRGβ1-induced BC growth”) and C. BT-474 and SK-BR-3 human BC cells treated with HRG β1 that represent a model comparable to JIMT-1 cells. Particularly, SK-BR-3 cells treated with HRGβ1 are a model closer to JIMT-1, since they lack estrogen and progesterone receptor (“Comparison between hErbB-2ΔNLS and Ttzm effects in ErbB-2 activation and BC growth”). The authors apologize for the mistake in the assembly of Figure 3f, which does not affect the conclusions of the article.