Reader Comments
Post a new comment on this article
Post Your Discussion Comment
Please follow our guidelines for comments and review our competing interests policy. Comments that do not conform to our guidelines will be promptly removed and the user account disabled. The following must be avoided:
- Remarks that could be interpreted as allegations of misconduct
- Unsupported assertions or statements
- Inflammatory or insulting language
Thank You!
Thank you for taking the time to flag this posting; we review flagged postings on a regular basis.
closeSeveral problems with study design and presentation
Posted by hookerthomasj on 04 Jul 2015 at 08:35 GMT
The authors aimed to examine the effect of primary human CRC-derived DNAs on HT29 colon adenocarcinoma and HDFalpha fibroblast cell lines.
However, there are several issues in this study that seems to be half-baked.
1. The authors used DNA isolated from human CRC samples. Regarding TLR-signaling examination, how did they ascertain that the isolated DNA used for incubation is protein, RNA and most importantly lipopolysaccharide free?
2. The authors examined only TLR9 in this study. Are they sure that other TLRs did not contribute to the results observed? How can the signaling effect of other TLRs be excluded?
3. In their previous study (Furi I et al, ScientificWorldJournal. 2013) the authors described that the biologic effect of DNA incubation is determined by the lenghts and methylation status of the used DNA fragments, which is a logical statement. In this study, the authors used "mixed" DNA samples without analysing the characteristics of the DNA fragments. How can they be sure that only pro-metastatic effects are present in the cells after DNA incubation? What kind of anti-metastatic, anti-inflammatory, pro-death genes were assayed during the whole gene expression assays? How did the expression of these genes alter after DNA incubation?
4. The strangest, not logical step in the study is the use of HDF-alpha cell line. The authors examined the effect of CRC-derived DNA on HT29 adenocarcinoma cells and Human Dermal Fibroblast (HDF) cells. How can CRC-derived DNA be in contact with dermal fibroblasts in vivo? Why did not they used normal fibroblast cell line from colonic origin (for example CCD-18co cell line) as normal colonic fibroblast? There are several genetic and epigenetic differences between colonic and dermal fibroblasts, so the results cannot be presented as relevant results regarding the pro-metastatic effect of CRC-derived DNA incubation, in my opinion.
5. Morover, histamine is one of the most important factor in dermal inflammation affecting dermal fibroblasts. How did the authors prove that the used DNA solution was free of histamine?
6. Another strange aspect of their result: they found that after DNA incubation the protein expression of both E-cadherin and cytokeratin 20 elevated. E-cadherin displays anti-apoptotic effects and favors sticking to the extracellular matrix, favoring the mesenchymal phenotype. While CK20 favors the epithelial phenotype by indicating differentiation. How can a cell show equal differentiation to both ways epithelial and mesenchymal phenotypes? In case of MET (mesenchymal-epithelial transition) the expression of E-cadharin increases, the expression of CK20 decreases. In case of EMT (epithelial-mesenchymal transition) the situation is the opposite. How can the authors explain their observation?
7. In the results the authors found alterations in the expression of DNMT1. However, they discussed the effect of DNMT3a in the discussion. The biologic function of DNMT1 and DNMT3a is not the same. This is also a serious mistake in their study.
I can not understand how the reviewers of this article did not take notice of these important aspects.
Thank you for reading my comments.
RE: Several problems with study design and presentation
belamolnar replied to hookerthomasj on 10 Jul 2015 at 11:56 GMT
Dear Dr. Hooker,
I would like to thank you for your comments about our manuscript. Hereby I would like to provide responses to the questions regarding our study design and results.
1. In the methods section we described that we added 20 μl RNase A/T1 Mix to tissue lysates and incubated at 37°C for 1 hour. According to the High Pure PCR template preparation kit manual proteinase K digestion(70 C, 30 min) was performed during DNA isolation.
2. We did not detected any remarkable gene expression alteration of LPS sensing receptors and other TLRs and TLR pathways using whole genome microarray analysis (logFc≥1; p≤0.05).
3. The application of tumour and healthy DNA was here introduced as the cells in vivo face similarly cancer or injured healthy DNA. This is a proceeding, enhacement as compared to our previous study, where methylated and non methylated DNA was applied only.
The application of whole-genome expression array aimed to determine for pro- and anti-metastatic effects of tumor and normal colon tissue derived DNA on HT-29 cells.
HGU 133 2.0 plus array is a suitable solution for this analysis, as it contains many pro- or anti- metastatic genes (e.g. MACC,MALAT MTA1, MTA3. NUPR1, MTSS1, BRMS1, CD82).
According to our results only MACC and MALAT 1 showed significant overexpression after colon cancer-derived tumor DNA treatment which supports our conclusion. From the genes affecting apoptosis only PPARD –as an anti-apoptotic gene- showed significant overexpression after tumor-derived DNA treatment supporting our conclusion. (logFc≥1; p≤0.05).
4. We used HDF-a cell line as a control cell line in order to determine the gene expression changes at the elements of TLR pathways after normal colon and colorectal tumor-originated DNA treatment. Simply we chould have access to fibroblasts from dermal origin at the initation of the study. Further expression analysis will be needed to determine the role of fibroblasts in colon carcinogenesis, as we in the last year we could establish cancer associated, non associated fibroblasts and even cancer epithelial cell cultures from the same patient. These studies are under writing and submission.
5. According to our results, no gene expression changes at histamine sensing pathways was detected after both, either normal or tumor DNA treatment (logFc≥1; p≤0.05). However, this is a very good idea to perform QC analysis of the DNA samples by MS before starting experiments to determine the level of possible contaminants including protein or polysaccharide components.
6. E-Cadherin mediates homotypic epithelial Ca dependent cell-cell adhesion and it is nothing to do with matrix adhesion. Its elevation is rather against EMT transition than behind it. Though it was elevated in treated cancer cells but rather in the intracytoplasmic than in the cell membrane compartment. Elevated cyokeratin 20 expression we can not explain it could be a rebound (compensatory) effect which may, however, also against EMT. Any in vitro system, our model too, suffers from the lack of the complex microenvironment around the growing cancer. Concernin EMT, obviously more appropriate conclusions could be drawn from an in vivo model.
7. In our previous study (Furi I et al, ScientificWorldJournal., 2013) we examined the expression level of all DNMTs. Relying on these observation in the present study we examined only the expression of DNMT3a. According to the fact that Affymetrix HGU 133 Plus 2.0 microarrays don’t include annotated DNMT3A probes, we further analysed DNMT1 expression levels, but no remarkable gene expression differences could be found between the control and the treated cells. This was the reason why we concentrated on DNMT3A in our immunocytochemistry (ICC) experiments.
We consider the sudy and its design an optimized, balanced one between aims and opportunities.
We further like to ask Your consideration for the major results of the study, namely, that the release and increase of cell free DNA contributes to a cancer cell survival and metastasis.
We will publish very soon further results and outcomes with other stromal cell types. Here CAFs, colon fibroblasts could have been involved, in addition to cell of hematopoietic origin.
Bela Molnar M.D., PhD.
as of Planning, Leading of this Study.
RE: RE: Several problems with study design and presentation
hookerthomasj replied to belamolnar on 14 Jul 2015 at 19:03 GMT
Dear Bela Molnar M.D., Ph.D.,
After reading your responses I still have doubts about the results of the published study. Here are my comments regarding your answers and explanations:
Ad 1. RNase and proteinase K treatment is not enough to ensure a solution to be lipopolysaccharide free, since lipopolysaccharides are not proteins or RNA structures.
It is still unclear how depyrogenation was carried out.
Ad 2. It is not enough to study the expression of TLRs, as there are other LPS sensing pathways in a living cellular model.
Ad 3. I understand that after performing whole genomic gene expression profiling several metastasis-associated genes showed expression alterations. But did you validate these results by conventional PCR, or more importantly by Western blotting?
Ad 4. The explanation namely "Simply we chould have access to fibroblasts from dermal origin at the initation of the study." is absolutely inadmissible. How can be dermal fibroblasts used for modelling the colonic pericryptal fibroblasts? In 2004, by using microarray expression analysis Nakagawa et al. proved that colonic fibroblasts clustered tightly into one group and skin fibroblasts into another. After careful reading of your article no words can be found about the origin of the used HDF alpha cell line, you just simply termed these cells as normal fibroblasts. No explanation about the misinterpretation of the results can be found in the article, even in the discussion. Instead of HDFalpha cells astrocytes also could be used for studying the intercellular context, but the expected results would not be relevant for modeling the colonic pericryptal fibroblast zone.
Ad 6. In your study you wrote that E-cadherin protein expression elevated after DNA treatment of HT29 cells. As E-cadherin is a known suppressor of tumor invasion and metastasis there exists a contradiction in your presented results, namely between the "pro-metastatic" gene expression profile vs. the immunocytochemical findings. That is why it is not acceptable to conclude on the basis of Affymetrix array results only.
Ad 7. According NetAffyx the Human Genome U133 Plus 2.0 Arrays you used in your study has the following DNMT oligo-sequences: DNMT1 (201697_s_at); DNMT3B (220668_s_at); DNMT3A (218457_s_at, 222640_at, 244428_at). So your explanation, namely "According to the fact that Affymetrix HGU 133 Plus 2.0 microarrays don’t include annotated DNMT3A probes, we further analysed DNMT1 expression levels..." is absolutely incorrect. There are several differencies between the biologic functions of DNMT1 and DNMT3A. On the other hand, your DNMT1 gene expression analysis was performed by qRT-PCR by using custom-designed primers. Why did not you design primers for DNMT3A?
In my opinion, your future plans cannot explain the scientific imperfection of the current study.
I am really surprised how the reviewer(s) of this study overlooked these important issues of this article.
Prof. T. Hooker
Professor of Oncology and Cellular Biology, ISREC