Interesting paper - I've noticed that exon coverage has been low in exosome RNA-Seq experiments, and it was interesting to see how much non-coding reads specifically aligned to rRNAs (although I expect there are also other issues in the data that I have worked with).
I have two suggestions that I think might be helpful to address when submitting the paper for peer review:
1) The abstract mentions that "depletion of fragmented rRNA should be utilized in the future RNA-seq analyses on exosomes", but I think a paper focusing on the differences in gene expression between exosomes should provide some data with rRNAs already depleted. It looks like you tried to use the RiboMinus Eukaryote kit, but it didn't work very well. I think this is a relatively common problem for small RNA-Seq, and you may want to research some alternatives (with RNase H being the most popular strategy, I believe). For example,you may want to consider some of the tools used in this paper:
2) Exosomes are a popular topic now, so I think it would help to add a comparison with publicly available data. I just did a quick search, and these look like a couple datasets that might be useful to compare:
I hope this helps - good luck!