You should validate reference genes experimentally in your conditions.

Despite raising awareness of potential pitfalls of qPCR, use of experimentally validated reference genes for data normalisation is still widely disregarded.

You will need to gather several miRNAs (>3) and use a normaliser method such as GeNORM/NormFinder.

I would recommend doing a literature search for miRNAs expressed in the exosomes you are looking at and select at least 4 miRNAs that will give you a robust normaliser. 

http://www.exosome-rna.com/identification-of-endogenous-controls-for-analyzing-serum-exosomal-mirna/

I know it is a tedious process but will make your data robust.

You should validate reference genes experimentally in your conditions.

Despite raising awareness of potential pitfalls of qPCR, use of experimentally validated reference genes for data normalisation is still widely disregarded.

You will need to gather several miRNAs (>3) and use a normaliser method such as GeNORM/NormFinder.

I would recommend doing a literature search for miRNAs expressed in the exosomes you are looking at and select at least 4 miRNAs that will give you a robust normaliser. 

http://www.exosome-rna.com/identification-of-endogenous-controls-for-analyzing-serum-exosomal-mirna/

I know it is a tedious process but will make your data robust.

U6 is certainly incorrect reference gene for exosomal miRNA qPCR because it is not in size range of miRNAs and is not encapsulated in exosomes as same as miRNAs. Use typic miRNAs such as miR-16 and other expressed miRNAs from cells secreted exosome (if known).

There are no miRNAs or other RNAs that are widely accepted to be good normalisers for exosomes. This hasn't stopped people picking ones at random and claiming they are normalisers though, so be extremely sceptical when you read the literature.

An alternative approach is absolute quantification. This is a lot more expensive and time consuming, but it does work.

Another aspect is to be aware what miRNAs are you actually probing and if they are real miRNAs or correctly annotated if they are real...check attached paper and MirGeneDB.org

Article A Uniform System for the Annotation of Vertebrate microRNA G...

There really are no specific miRNAs identified that are consistently expressed in exosomes, and exosomal miRs can vary by cell type. The best alternative is to use a spike-in control, there are several available. Qiagen uses C.elegans miR-39. 

Dear David

I agree with Elizabeth, see an example of results in my lab where we used exoRNeasy Serum/Plasma Midi Kit and found more variation in miR-16 than in cel-miR-39.

Wach the .ppt file.

In my opinion both U5 and U6 snRNAs work as references. For insects, miR-184 is usually abundant in all tested tissues, thus I guess it is a potential candidate. I also agree that the use of spike-in control is a very good alternative.