Identification of Mispairing Omic Signatures in Chinese Hamster Ovary (CHO) Cells Producing a Tri-Specific Antibody

Round 1

Reviewer 1 Report

In this manuscript, the authors described a novel approach to detect presence/absence of mispairings in MsAbs grown in CHO cells. In my opinion, this is an innovative method, although I have some comments that should be considered:

1. Figure 3 and 5. When plotting volcano plots and doing pairwise comparisons, the authors should use adjusted p-values for their calculations. This is to minimise the risk of false discovery since the authors are probably measuring many different transcripts simultaneously using the RNAseq. In addition, some of the interesting genes (especially at the edges of the volcano plot) can be annotated.

2. Figure 3 and 5. Besides doing a differential analysis based on control cells, the authors should also plot one that shows the differential analysis between high and low pairing clones. In my opinion, this is even more critical than the comparison to the control cells since we are interested in finding genes that separate these 2 conditions the most.

3. Besides volcano plot, having a MA plot for the RNAseq results will show the abundance of the genes measured. Sometimes this is useful as some genes have great effect size, but attributed to low abundance counts. For instance, 1 vs 10 count and 10 vs 100 counts are both 10-fold increase, but the 10 vs 100 counts are more interesting because the readings for 10 counts is more accurate than 1 count.

4. Pathways should highlight adjusted p-values as well, and consider direct comparisons between high and low pairing clones.

5. Figure 6. The authors should plot the ROC curves for the different transcripts and with all genes combined to test the sensitivity and specificity of the genes in accurately predicting the high and low pairing clones

Author Response

Summary: We thank the reviewer for taking the time to review this manuscript. Please find below the point-by-point answers to your comments. The corresponding revisions are highlighted in yellow in the re-submitted files.

Comment 1: Figure 3 and 5. When plotting volcano plots and doing pairwise comparisons, the authors should use adjusted p-values for their calculations. This is to minimise the risk of false discovery since the authors are probably measuring many different transcripts simultaneously using the RNAseq. In addition, some of the interesting genes (especially at the edges of the volcano plot) can be annotated.

Response 1: We thank the reviewer for the suggestion. Since we only include significantly regulated genes/proteins in our pathway enrichment analysis (and not all the quantified transcripts, avoiding introducing biased data), we decided to keep p-value ≤0.05 as our threshold, in order to include a higher number of DEG genes/proteins representative of the different cellular pathways and obtain a better overview in this functional enrichment analysis. In line with the reviewer's comment, we have now added the adjusted p values for all analysis performed, in the supplementary files (Sup files 3, 4, 5, and 6).

We have also annotated the genes/proteins with the highest -log(p-value) and highest/lowest log2(FC) in the volcano plots (fig 3 and 5), as suggested by the reviewer.

Comment 2: Figure 3 and 5. Besides doing a differential analysis based on control cells, the authors should also plot one that shows the differential analysis between high and low pairing clones. In my opinion, this is even more critical than the comparison to the control cells since we are interested in finding genes that separate these 2 conditions the most.

Response 2: We acknowledge the reviewer for the suggestion.  We agree that the comparison between high and low mispairing clones is indeed the most critical. It was based on this that we presented the differential analysis using the volcano plots on Figures 3 and 5 which correspond to the comparison between high mispairing clones group and low mispairing clones group, as indicated in the legend “Differential gene and protein expression analysis between low and high mispairing clones…”. There are no control cells in this analysis, instead, there are two groups of clones (supplementary figure 2): clones with low mispairing levels (percentage of correct tsAb mass ≥ 90%) and clones with high mispairing levels (percentage of correct tsAb mass < 90%).

Comment 3: Besides volcano plot, having a MA plot for the RNAseq results will show the abundance of the genes measured. Sometimes this is useful as some genes have great effect size, but attributed to low abundance counts. For instance, 1 vs 10 count and 10 vs 100 counts are both 10-fold increase, but the 10 vs 100 counts are more interesting because the readings for 10 counts is more accurate than 1 count.

Response 3: We thank the reviewer for this suggestion. We now added MA plots for transcriptomics datasets at day 5 (exponential phase) and day 10 (production phase) in supplementary figure 6. As depicted by a horizontal black line in the graphs, only genes with base mean values equal or higher than 1 were analysed.

Comment 4: Pathways should highlight adjusted p-values as well, and consider direct comparisons between high and low pairing clones.

Response 4: We appreciate the reviewer comment. As stated in the methods section, “Statistically significant representation of biological functions and canonical pathways was identified based on IPA p-value, displayed as −log (p-value). This probability score is calculated taking into account the total number of molecules known to be associated with a given function or pathway, and their representation in the experimental dataset.”. There is no adjusted p-values computed by IPA (Ingenuity Pathway Analysis) software. As mentioned above, we give as input to IPA only DEG with p value ≤0.05 from the direct comparison between high and low mispairing clones.

Comment 5: Figure 6. The authors should plot the ROC curves for the different transcripts and with all genes combined to test the sensitivity and specificity of the genes in accurately predicting the high and low pairing clones.

Response 5: We thank the reviewer for the suggestion. We did not apply any predictive model in this paper, therefore, no ROC curve analysis are displayed. In figure 6, our main goal is to use PCR analysis to validate RNAseq data. For that we selected a panel of genes that are differentially regulated between the two groups of clones analysed.  We do agree with the reviewer that it would be interesting to further explore a subset of transcripts to be used and validated as a potential mispairing biomarker panel, as stated in the last paragraph of the discussion “(…) the development of a biomarker panel that could be applied as an additional tool for early screening and selection of clones with more suitable product profiles”. However, this is out of scope for this publication.

Reviewer 2 Report

In this study, the Authors have performed quantitative transcriptomics and proteomics analyses to investigate which signalling pathways correlate with low and high mis-pairing clone signatures. Gene and protein expression profiles of Chinese Hamster Ovary clones producing an tsAb were analysed in exponential growth and stationary phase of fed-batch culture. The concept of the study is interesting and scientifically important. 

Specific comments:

1. Description of cell culture conditions, media etc should be extended.

2. For clarity, please prepare a table with sequences of primers used in PCR experiments.

3. The Authors should discuss the limitations of their study.

4. The discussion should be slightly modified, and the repetition of the results should be avoided. Especially, it is not necessary to refer to specific figures in the discussion.

5. In supplementary files, I cannot find figures. The Authors refer to Figures S... If they mean Excel datasheets, which are uploaded, the names of this supplementary materials must be changed to e.g., "File S1" etc. 

6. Please correct typos and formatting.

Author Response

Summary: We thank the reviewer for taking the time to review this manuscript. Please find below the point-by-point answers to the comments. The corresponding revisions are highlighted in yellow in the re-submitted files.

Comment1: Description of cell culture conditions, media etc should be extended.

Response 1: We thank the reviewer for this suggestion. We have added some details on initial cell density, temperature, target pH and oxygen percentage: “To ensure consistency between clones, the same process parameters were applied, including initial cell density (1E6 viable cells/mL), temperature (36,5°C), feeding regime, oxygen (40% DO) and pH control (target pH 7).”

Medium, feed, and feeding regimen used cannot be disclaimed due to legal restrictions, as stated in Data Availability Statement paragraph.

Comment 2: For clarity, please prepare a table with sequences of primers used in PCR experiments.

Response 2: Following the reviewer’s suggestion, we have replaced, in the methods section “Validation of Differential gene expression by RT-PCR”, the paragraphs describing the primers sequences used by a table (table 1).

Comment 3: The Authors should discuss the limitations of their study.

Response 3: We completely agree with the reviewer that it is always important to state the study limitations. We believe the main limitation of this work is the small number of samples, as well as a lack of balanced distribution of antibody productivity between the two groups analysed, as we state in the last paragraph of the discussion, in the following sentence: “While the small cohort of samples used in this study, and the unbalanced distribution of productivity levels by the groups analysed could impact the statistical power of the findings, (…)”.

Comment 4: The discussion should be slightly modified, and the repetition of the results should be avoided. Especially, it is not necessary to refer to specific figures in the discussion.

Response 4: We thank the reviewer for this comment. We have removed the references to the figures in the discussion section.

Comment 5: In supplementary files, I cannot find figures. The Authors refer to Figures S... If they mean Excel datasheets, which are uploaded, the names of this supplementary materials must be changed to e.g., "File S1" etc. 

Response 5: We thank the reviewer for the note. There are supplementary files as well as supplementary figures in this manuscript.

The supplementary figures were included in the submission, in the zip folder corresponding to the figures, as this was our understanding when reading the author guidelines.  On this re-submission, we have also added the supplementary figures to the zip folder corresponding to the supplementary information.

Comment 6: Please correct typos and formatting.

Response 6: We thank the reviewer for the suggestion. All the typos detected have been corrected.

Round 2

Reviewer 1 Report

Authors have sufficiently answered to my comments