On 2015 Aug 16, Xiang Ming commented:

Thanks very much for Murtaugh’s careful review. According to your comment, we make following explanations. 1. The pancreas photos on the first row of Fig.5A showed that the surgical removal of tissues do not include intestine. After being photographed, the tissues were embedded in paraffin, so we could ensure that all sections assessed in this paper are from pancreas. 2. As you said “the treatment might have induced cancers of an intestine-like pathology”, we should explain that a great quantity of non-tumor components existed in pancreatic cancer including stromal cells and lymphocytes. The immunohistochemical analysis in Figure.5 was consistent with this pathological characteristic which was similar to intestine-like pathology. 3. In our study, the malignant cell could secrete a small amount of amylase (Fig.5A). This also prove that the treatment didn’t induce cancers of an intestine-like pathology. On the other hand, the Ki67 is mainly expressed in the malignant acinar cell nucleus, so high levels of Ki67 expression could support our conclusion that Reg3g could promote proliferation in acinar cells

On 2015 Jul 30, L Charles Murtaugh commented:

While this is an interesting study, and the role of REG genes in pancreatic cancer is arguably understudied, the conclusions of this paper are tempered by irregularities in the histological analysis. Only one study group developed pancreatic cancer in this study, namely the high-dose pReg3g + DMBA treatment, referred to as HA10R. Inspection of the histology data presented for the HA10R group (Fig. 5A) reveals that several of the most relevant images (H&E staining, Ki67 and cytokeratin-19), for this group specifically, appear to be taken from sections of the intestine rather than the pancreas. While, in principal, the treatment might have induced cancers of an intestine-like pathology, this would not explain the clear organization of proliferative crypts and non-dividing villi, apparent from Ki67 staining. This raises doubts about the quantitative analysis of these and other variables, in Fig. 5B, as well as about the overall conclusion that treated mice developed cancer of the pancreas.

On 2015 Jul 30, L Charles Murtaugh commented:

While this is an interesting study, and the role of REG genes in pancreatic cancer is arguably understudied, the conclusions of this paper are tempered by irregularities in the histological analysis. Only one study group developed pancreatic cancer in this study, namely the high-dose pReg3g + DMBA treatment, referred to as HA10R. Inspection of the histology data presented for the HA10R group (Fig. 5A) reveals that several of the most relevant images (H&E staining, Ki67 and cytokeratin-19), for this group specifically, appear to be taken from sections of the intestine rather than the pancreas. While, in principal, the treatment might have induced cancers of an intestine-like pathology, this would not explain the clear organization of proliferative crypts and non-dividing villi, apparent from Ki67 staining. This raises doubts about the quantitative analysis of these and other variables, in Fig. 5B, as well as about the overall conclusion that treated mice developed cancer of the pancreas.