Development was made aware by a reader of potential duplication of data in Fig. 3 and Fig. S3 of Development (2014) 141, (doi:10.1242/dev.106492).
The journal contacted the authors who said that some of the bands in the western blots in Fig. 3 and Fig. S3 were manipulated during figure compilation. After discussion with the corresponding author, Development referred this matter to Icahn School of Medicine at Mount Sinai (ISMMS), who investigated and concluded that: “The main conclusions of the manuscript are supported by the data obtained utilizing other methodologies, including immunohistochemistry.” Development's editorial policies state that: “Should an error appear in a published article that affects scientific meaning or author credibility but does not affect the overall results and conclusions of the paper, our policy is to publish a Correction…” and that a Retraction should be published when “…a published paper contain[s] one or more significant errors or inaccuracies that change the overall results and conclusions of a paper…”. The journal follows the guidelines of the Committee on Publication Ethics (COPE), which state: “Retraction should usually be reserved for publications that are so seriously flawed (for whatever reason) that their findings or conclusions should not be relied upon”.
The standards of figure assembly and data presentation in this paper fall short of good scientific practice. However, given that the ISMMS declared that the conclusions of the paper were not affected by the manipulations, the appropriate course of action – according to COPE guidelines – is to publish a Correction, which Development has made as detailed as possible. The corresponding author has provided the journal with the original scans of the relevant blots.
In the Mbp blot in Fig. 3C, lanes were rearranged to match the order of samples on the other blots without an appropriate explanation. These splices are now indicated with black lines in the new figure panel. The control actin blot in Fig. 3C was compressed vertically when the figure was prepared. The uncompressed blot is shown below, along with a new densitometric analysis in Fig. 3D based on these bands.
Tgfβ1, ActB and co-treatment elicit distinct signaling and functional outcomes in OLPs. (C,D) Immunoblotting (C) and densitometry (D) of primary OLPs plated into serum-free media and exposed to 50 ng/ml Tgfβ1, ActB and/or Bmp4 for 5 days. Data accompany morphometric timecourse analyses in OLPs in E-J and supplementary material Fig. S4E-G. At 5 days, caspase-3 cleavage (a marker of apoptotic activity) was decreased in OLP cultures exposed to Tgfβ or ActB alone or together. Mbp expression (mature oligodendrocytes) was enhanced in ActB-treated and co-treated cultures. Bmp alone or in combination with Tgfβ or ActB abrogated Mbp expression.
Tgfβ1, ActB and co-treatment elicit distinct signaling and functional outcomes in OLPs. (C,D) Immunoblotting (C) and densitometry (D) of primary OLPs plated into serum-free media and exposed to 50 ng/ml Tgfβ1, ActB and/or Bmp4 for 5 days. Data accompany morphometric timecourse analyses in OLPs in E-J and supplementary material Fig. S4E-G. At 5 days, caspase-3 cleavage (a marker of apoptotic activity) was decreased in OLP cultures exposed to Tgfβ or ActB alone or together. Mbp expression (mature oligodendrocytes) was enhanced in ActB-treated and co-treated cultures. Bmp alone or in combination with Tgfβ or ActB abrogated Mbp expression.
The order of the lanes was incorrect in the P-GSK3α/β(S21/S9) blot in Fig. S3B. The correct order is shown in the new figure below, with lines indicating where lanes have been spliced. The investigation also showed that bands from the original P-GSK3α/β(S21/S9) blot were rearranged and duplicated in the actin blot. The correct actin blot for Fig. S3B was located and is shown in the new figure. Readers should note that the P-p44/42(T202/Y204) blot in Fig. S3B has also been replaced because the original blot was inappropriately over-contrasted. On the basis of these changes, a new densitometric analysis was carried out and is reported in Fig. S3C-J. In addition, the actin loading control blot in Fig. S3K was duplicated from Fig. 3A. The correct actin blot for Fig. S3K was located and is shown in the new figure below.
Tgfβ1, ActB and co-treatment of Oli-Neu cells elicit distinct patterns of canonical Smad-dependent and non-canonical MAP kinase signaling. (A) Primary rat OLP cultures grown in medium favoring proliferation and non-permissive for differentiation were fixed, immunolabeled for Tgfβ superfamily ligand-binding receptors and Olig2, and imaged by confocal microscopy. Cells were ubiquitously positive for Tgfβr2, ActrIIb and Bmpr2. (B-J) Immunoblotting (B) and densitometric data (C-J) from Oli-Neu cells plated into serum-free media and exposed to 50ng/ml Tgfβ1 and/or ActB for 15, 30 or 60min. Findings complement data from primary OLP in Figs.3A and 3B. (B,C) Tgfβ1 or ActB alone each induced Smad3 phosphorylation at Ser423/425, although the effect of Tgfβ1 was lost by 60min. However, the effect of ActB was stronger than the equivalent Tgfβ1 concentration at all three timepoints, and this difference became more pronounced over time. In contrast, Tgfβ1 activation of p42/44 MAP kinase (P-Thr202/Tyr204) was seen at 15min, whereas ActB treatment persistently reduced p42/44 phosphorylation (B,D). Notably, co-treatment with Tgfβ1 plus ActB together produced a third distinct pattern, which combined increased levels of P-Smad3 (Ser423/425) similar to or beyond those induced by ActB alone, with increased phosphorylation of p42/44 MAP kinase beyond that induced by Tgfβ1 alone at later timepoints (B-D). No significant changes were seen in levels of total Smad3 or p42/44 proteins (B,E,F). Neither ligand alone or in combination impacted activity of the Akt-Gsk3 signaling pathway, as measured by changes in levels of phosphorylated Akt (P-Thr308) or Gsk3α/β (P-Ser21/9) (B,G-I). Smad3 phosphorylation within its linker region at Ser208 resulting from non-canonical pathway activation has been shown to alter its transcriptional activity, but no changes in Smad3 (P-Ser208) were detected in Tgfβ1- or ActB-treated cultures (B,J). (K,L) Oli-Neu cultures plated into serum-free media were exposed to 50ng/ml Tgfβ1, ActB or vehicle for 30min, then were subjected to co-immunoprecipitation using anti-Smad3 antibody or IgG control. HEK cells were used for comparison of cell type specificity. Immunoprecipitates and lysates were then subjected to immunoblotting and densitometric analysis. Blots of immunoprecipitates were probed for potential Smad3-interacting coactivators, and lysate blots were probed for actin (loading control). In Oli-Neu cells (but not HEK cells), Smad3 was found to associate with FoxH1/FAST (K), but not with other potential Smad-interacting factors, including ETF, Sp-1, Gli, TCF, or FoxO1. No differences in Smad3 binding to FoxH1/FAST were observed following ActB or Tgfβ1 treatment (K,L). Data are representative of findings from 3 independent experiments in separate cultures. Scale bars, 10 µm (A).
Tgfβ1, ActB and co-treatment of Oli-Neu cells elicit distinct patterns of canonical Smad-dependent and non-canonical MAP kinase signaling. (A) Primary rat OLP cultures grown in medium favoring proliferation and non-permissive for differentiation were fixed, immunolabeled for Tgfβ superfamily ligand-binding receptors and Olig2, and imaged by confocal microscopy. Cells were ubiquitously positive for Tgfβr2, ActrIIb and Bmpr2. (B-J) Immunoblotting (B) and densitometric data (C-J) from Oli-Neu cells plated into serum-free media and exposed to 50ng/ml Tgfβ1 and/or ActB for 15, 30 or 60min. Findings complement data from primary OLP in Figs.3A and 3B. (B,C) Tgfβ1 or ActB alone each induced Smad3 phosphorylation at Ser423/425, although the effect of Tgfβ1 was lost by 60min. However, the effect of ActB was stronger than the equivalent Tgfβ1 concentration at all three timepoints, and this difference became more pronounced over time. In contrast, Tgfβ1 activation of p42/44 MAP kinase (P-Thr202/Tyr204) was seen at 15min, whereas ActB treatment persistently reduced p42/44 phosphorylation (B,D). Notably, co-treatment with Tgfβ1 plus ActB together produced a third distinct pattern, which combined increased levels of P-Smad3 (Ser423/425) similar to or beyond those induced by ActB alone, with increased phosphorylation of p42/44 MAP kinase beyond that induced by Tgfβ1 alone at later timepoints (B-D). No significant changes were seen in levels of total Smad3 or p42/44 proteins (B,E,F). Neither ligand alone or in combination impacted activity of the Akt-Gsk3 signaling pathway, as measured by changes in levels of phosphorylated Akt (P-Thr308) or Gsk3α/β (P-Ser21/9) (B,G-I). Smad3 phosphorylation within its linker region at Ser208 resulting from non-canonical pathway activation has been shown to alter its transcriptional activity, but no changes in Smad3 (P-Ser208) were detected in Tgfβ1- or ActB-treated cultures (B,J). (K,L) Oli-Neu cultures plated into serum-free media were exposed to 50ng/ml Tgfβ1, ActB or vehicle for 30min, then were subjected to co-immunoprecipitation using anti-Smad3 antibody or IgG control. HEK cells were used for comparison of cell type specificity. Immunoprecipitates and lysates were then subjected to immunoblotting and densitometric analysis. Blots of immunoprecipitates were probed for potential Smad3-interacting coactivators, and lysate blots were probed for actin (loading control). In Oli-Neu cells (but not HEK cells), Smad3 was found to associate with FoxH1/FAST (K), but not with other potential Smad-interacting factors, including ETF, Sp-1, Gli, TCF, or FoxO1. No differences in Smad3 binding to FoxH1/FAST were observed following ActB or Tgfβ1 treatment (K,L). Data are representative of findings from 3 independent experiments in separate cultures. Scale bars, 10 µm (A).
The authors apologise to the journal and readers for these errors.
