Thank you for your comment. Our article reports the results of the first study on the analysis of RNA content of exosomes. Several previous studies also analyzed exosomal RNA using RNA-Seq approach but they focused on short RNAs (miRNAs, snoRNAs etc.). Therefore we tried a standard protocol for exosomal RNA preparation.
The unexpected conclusion of our study was that the vast majority of exosomal sequencing reads mapped to ribosomal RNAs. This was totally unexpected since Bioanalyzer results demonstrated absence (or near absence) of 18S and 28S peaks in samples of exosomal RNA. The result suggests that exosomes are packed mainly with rRNA fragments. Therefore, the answer to your question "do you think the current study justifies use of this protocol for biomarker applications?" is "not". We do not think that the used protocol is suitable for the biomarkers discovery. But that became clear only after the current study has been completed. What we proposed in the paper is to introduce a special protocol for the analysis of exosomal RNA that would include depletion of rRNA fragments. We demonstrated that recommended by Life Technology RiboMinus Kit was inefficient in removal of rRNA fragments. Hence, other approaches should be utilized in the future.
In regards with your question that "most of the genes selected for qPCR validation show high expression levels in several normal tissue types in BioGPS", I agree with it. When we conducted qPCR experiments our aim was to simply validate the RNA-Seq data irrespectively of the levels or uniqueness of RNA expression. Should your comment be made during the manuscript revision process we would be happy to test by qPCR the presence in exosomes of RNAs specific for breast cancer cells.
Regarding the question that "Figure 7 shows that most genes expressed in both cellular and exosomal MDA-MB-231 samples are also expressed in MDA-MB-436 exosomes. Do you think that it is safe to assume that genes uniquely expressed in MDA-MB-231 exosomes are also uniquely expressed in MDA-MB-231 cellular samples (in comparison to MDA-MB-436 exosomes and cellular samples, respectively)?", I can say that might be the case. However, the low number of reads mapped to non-rRNA sequence provided too little information to spot those cases. Again, this study should be repeated after depletion of fragmented rRNA (that is what we are doing now).
Regarding the question "do you think there will be differences between cell line exosome extraction and patient exosome extraction (in regards to protocol and/or resulting RNA quantities)?" my answer would be "yes and not". Regarding the protocol, it will be definitely different. Patients exosome extraction from the blood plasma, for example, would be done differently from exosome extraction from the cell conditioned medium. Human blood plasma contains up to 10e12 vesicles/ml, only a small fraction of which is represented by exosomes secreted by tumor tissue. However, in cell culture all exosomes are originated from cancer cells (if the medium does not contain FBS). In fact, we use different protocols for extracting exosomes from the plasma and cell culture medium.
Regarding the RNA quantity and quality. I expect that RNA quantity from blood plasma exosomes would depend on the method of getting specifically "cancer" exosomes and also on the size and type of tumor that sheds exosomes. Regarding the secreted RNA species, some of them might be common between those found in patients samples and cell culture. The following study, for example, reports that miRNAs released in exosomes from MCF7 breast cancer cell line are same that found in exosomes purified from the ductal lavages of breast cancer patients: Pigati L, et al. Selective release of microRNA species from normal and malignant mammary epithelial cells. PLoS One 2010;5:e13515