That's a tough question, and knowledge in the field is evolving rapidly... For exosomes (defined as small extracellular vesicles that are genuinely derived from late endosomes), a good start would be to assess syntenin or TSG101. We now know that the tetraspanins are way less specific to exosomes than initially believed. You could try to look for both CD81 and CD63; in combination these might also indicate the presence of exosomes.
Here's a great paper on the proteomic heterogeneity of EV populations:
Article Proteomic comparison defines novel markers to characterize h...