Author's response to reviews Title: Production and characterisation of monoclonal antibodies against RAI3 and its expression in human breast cancer Authors: Hannah Jörißen (hannah.joerissen@molbiotech.rwth-aachen.de) Nuran Bektas (nbektas@ukaachen.de) Edgar Dahl (edahl@ukaachen.de) Arndt Hartmann (arndt.hartmann@uk-erlangen.de) Anette ten Haaf (atenhaaf@ukaachen.de) Stefano Di Fiore (difiore@molbiotech.rwth-aachen.de) Hans Kiefer (hans.kiefer@m-fold.com) Andreas Thess (andreas.thess@curevac.com) Stefan Barth (stefan.barth@ime.fraunhofer.de) Torsten Klockenbring (klockenbring@molbiotech.rwth-aachen.de) Version: 2 Date: 17 April 2009 Author's response to reviews: see over
Prof. Dr. Dr. S. Barth UK Aachen EMI Pauwelsstr. 20 52074 Aachen To: BMC Cancer Editors Universitätsklinikum der RWTH Aachen Abteilung für Experimentelle Medizin und Immuntherapie Leitung Prof. Dr. Dr. Stefan Barth c/o Helmholtz-Institut für Biomedizinische Technik Pauwelsstr. 20 52074 Aachen - Tel.: +49-(0)241-80.80653 FAX: +49-(0)241-80.82442 E-Mail: barth@hia.rwth-aachen.de Dear Editors, Friday, 17 April 2009 - We thank you for giving us the opportunity to submit a revised version of the manuscript and we are grateful to the reviewers for their advice and support helping us to increase the quality of our manuscript. We have addressed the editorial revisions as requested. In the abstract, Results and Discussion as been changed to Results and the manuscript has been professionally copyedited and reformatted according to the journal style. In addition, we would like to note that we agree with considering our manuscript as a technical advance rather than a research article. The point by point answers to the reviewers' comments explaining and discussing the changes in the manuscript are outlined below. Please do not hesitate to contact us if you have any further questions. Yours sincerely Prof. Dr. Dr. Stefan Barth Prof. Dr. Dr. Stefan Barth Page 1 of 8 17.04.2009
Reviewer 1: Marcello Maggiolini Reviewer's report: In this study, Jörißen et al. reported the generation of a monoclonal antibody that detects RAI3 in formaldehyde-fixed paraffin-embedded tissue and a systematic analysis of RAI3 expression in normal and cancerous human breast tissue. The study mainly focuses on the RAI3 antibody production which would be more appropriate for a methodology targeted journal. The other aspect of the study although interesting, describes only observational data relative to RAI3 expression which should be extended to a functional investigation in order to clarify previous results shown by other Authors and to better assess the role exerted by RAI3 in breast cancer. We agree with the reviewer that extended functional investigations of RAI3 would be of special interest to clarify the role of RAI3 in breast cancer. However, the intent of this study was to generate monoclonal antibodies against RAI3 and to analyse RAI3 protein expression in a breast cancer TMA. Still, this antibody might be useful for further experiments addressing the functional role of RAI3. Therefore, we agree with the editors suggestion to consider the manuscript as a technical advance rather than a research paper. Reviewer 2: Nicolas de Roux With respect to the major compulsory revisions: Supplementary informations Comment 1: In RAI3 production, the authors used a RAI3 cdna cloned in frame with GST and His tag, produced inclusion bodies, cleaved the GST tag with thrombin, and purified the recombinant RAI3 using Ni-NTA superflow in presence of Glutathione. The authors must explain why they used a plasmid expressing RAI3 Prof. Dr. Dr. Stefan Barth Page 2 of 8 17.04.2009
in frame with GST? Why they did not use a recombinant RAI3 in frame with His Tag only. The attempt to produce eukaryotic membrane proteins in E. coli regularly leads to a toxic effect resulting in very low levels of recombinant membrane protein. Toxicity is probably a consequence of inefficient membrane insertion of non-bacterial protein sequences, leading to interference with the cellular components conveying insertion. The N-terminal GST fusion stops proteins from interacting with the membrane and directs them to inclusion bodies, thereby abolishing toxicity. As a result, yields of recombinant protein rise by several orders of magnitude. The above has been described in the cited reference [10]. A short explanation has been added to the methods section. Supplementary Controls: Comment 2: The authors used competitive-elisa to affirm specificity of their antibody and then expressed RAI3 in HEK293 to perform western blot. Western blot did not reveal any band in untransfected HEK293 cells. An additional control by immunohistochemistry must be performed on transfected and untransfected HEK293 cells. It will be interesting to perform this control in permeabilized and unpermeabilized cells. As additional control we have performed immunocytochemistry of transfected and untransfected HEK293T cells. We used cells transiently transfected with RAI3 in fusion to GFP allowing the direct detection of recombinant protein in the cell by fluorescence microscopy. We specifically detected recombinant protein in the transfected cells with the anti-rai3 antibody 24 2.3 and fluorescent-labelled Prof. Dr. Dr. Stefan Barth Page 3 of 8 17.04.2009
secondary antibody and analysed staining by laser scanning confocal microscopy. We could demonstrate distinct co-localisation of the detected signal with the recombinant RAI3-GFP emphasising specificity of the antibody, whereas no signal could be detected in untransfected cells. As we believe that this experiment is very valuable in demonstration of antibody specificity, we have added this experiment to the manuscript and generated the new Figure 2. Comment 3: Why the authors did not found any band in HEK293 untransfected cells whereas in introduction they reported a paper indicating that sirna against RAI3 reduces HEK293 cell growth (ref 5). This discrepancy with literature results must be explained or discussed. This is an interesting point to discuss. To further investigate this issue, we have performed additional western blots with available cell lines analysed in ref 5. The cell line MCF-7 is described as positive for RAI3 whereas levels in MDA-MB-546 were neglible. We could demonstrate endogenous RAI3 in MCF-7 cells but not in MDA- MB-546 correlating with the expression analysis of ref.5. However, no RAI3 could be detected in HEK293T wt cells. The according figure is provided as supplementary figure 1. In addition, on our immunocytochemistry experiment endogenous RAI3 could not be detected in untransfected HEK293T cell. It seems that levels of endogenous RAI3 in our HEK293T cells are below the detection threshold. One explanation may be that the HEK293 cell lines used have clonal differences that lead to this discrepancy. This is a problem in general as there is a tremendous number of different factors that could account for changes of the physiological status of cells and can affect the expression levels of certain proteins, e.g. age of the cells, numbers of passages, use of cell culture medium, supplements, etc. We have added this point Prof. Dr. Dr. Stefan Barth Page 4 of 8 17.04.2009
to the discussion section and added an additional reference addressing this issue [Ref 24]. Comment 4: In the western blot, they explained their multiple bands by RAI3 proteolysis. It could be an explanation, however, cross hybridization with other molecules is another explanation. It is quite easy to test this hypothesis by performing sirna experiment in breast cancer cell line expressing RAI3. These controls are absolutely necessary before performing any quantitative analysis of RAI3 expression using tissue microarray. We agree with the reviewer that sirna analysis would be a good method to show RAI3 specificity of the antibodies. However, in the case that the antibody would bind to cross-hybridizing proteins, these unspecific bands would also appear in the mock transfected cell lysate. In addition, the clear shift of the whole band pattern when using lysates of RAI3-GFP transfected cells indicates specificity for RAI3. Taking together the western blot data and our additional experiments including the demonstration of co-localisation of RAI3-GFP and the signal detected with anti-rai3 antibodies and the competitive IHC control (see also below) we believe that specificity of the antibody is adequately demonstrated. Comment 5: The authors has performed cdna dot blot to analyse RAI3 expression. It is an old technique for such purpose. Why they did not use quantitative RT-PCR which is clearly more accurate? The reviewer is right that the cdna dot blot technique is a quite old technique. However, the Clontech Cancer Profiling Array has two advantages compared to real Prof. Dr. Dr. Stefan Barth Page 5 of 8 17.04.2009
time PCR: a) This standardized cdna format has been used in numerous studies making it possible to better compare expression data of cancer related genes; b) the Cancer Profiling Array contains tissue derived from 100 tissues (50 breast tumour and 50 corresponding normal breast tissues), the total number and especially the number of matching normal tissue is very difficult to get from fresh frozen tissues. Comment 6: They don t show any northern blot on total RNA extracted from breast cancer? How they analysed specificity of cdna probe? We have performed a Northern Blot (see supplementary Figure 2) using the Clontech Multiple Tissue Northern Blot (MTN) containing poly A+ RNA derived from human heart, brain, placenta, lung, liver, skeletal muscle, kidney and pancreas. This blot detects two known RAI3 transcripts of 6.8 and 2.4 kb in size and demonstrates extraordinary abundant RAI3 expression in human lung tissue. This observation has already been published by Nagahata et al. (Endocr Relat Cancer 2005, 12:65-73) (their Figure 3) confirming the specificity of our cdna probe. Since the information on predominant lung expression has been published, we suggest not to include this Northern Blot as an additional figure in the manuscript. Unfortunately, none of the Clontech Northern blots contains RNA derived from human normal breast tissue. Prof. Dr. Dr. Stefan Barth Page 6 of 8 17.04.2009
Comment 7: Did they use a full length cdna or only a portion? They must precise negative controls used in this experiment. We have to apologise that we indeed did not describe precisely the RAI3 cdna probe used for Northern blot and cdna dot blot hybridization. The RAI3 cdna probe used included base pair 550 to 2835 of the RAI3 cdna deposited under Accession Number NM_003979. We have included this information in the text. Negative controls are always contained on the Cancer Profiling Array. In Supplementary Figure 3 we show the complete array with negative controls (and cell lines) that are positioned at the right side of the nylon membrane. This analysis shows that yeast total RNA, yeast trna, E.coli DNA, poly A+ RNA and ubiquitin cdna do not exhibit a cross-hybridization signal. Genomic DNA of course contains the RAI3 gene, therefore a weak hybridization signal can be expected. Finally the probe was hybridized to an internal positive control (spotted RAI3-cDNA) and was able to detect less than 10 pg of RAI3 cdna. Comment 8: Immunohistochemistry analysis on breast cancer is not convincing. Why they use a non-specific negative control by omitting the first antibody. When immunogen is available, the true negative control is the immunoneutralisation of the primary antibody. This analysis must be performed. We apologise that we have not included such a negative control in the manuscript so far. We performed unspecific negative control to show that there was no unspecific background staining from secondary antibodies. However, we have now performed an additional control by competitive immunohistochemistry on a breast cancer Prof. Dr. Dr. Stefan Barth Page 7 of 8 17.04.2009
sample showing the suppression of RAI3 staining by preincubating the Anti-RAI3 Mab 24 2.3 with 200-fold molar excess of recombinant RAI3. The experiment is added as figure 5 in the manuscript. Comment 9: Without these additional controls, results expressed in table 2 are not relevant and must be removed from the manuscript. We have performed or discussed the suggested controls and therefore believe that the results presented in table 2 should be included in the manuscript. Minor Essential Revisions 1: Page 19 : transcription of mrna should be read traduction. The reviewer is right. We have changed transcription to translation (Page 21) 2: Needs some language corrections before being Published The manuscript has been professionally copy-edited. Prof. Dr. Dr. Stefan Barth Page 8 of 8 17.04.2009