Counting exosome secretion

Last month I wrote a post about exosome internalization by recipient cells.  One of the topics I discussed was the lack of good quantitative data in the exosomal field, and what the current data tells us about the efficiency and capacity of exosome-mediate cell-to-cell communiation.

Today I came across an interesting paper in which the researchers try to get quantitative data of exosome secretion by the donor cells.

The paper, published in the journal “Small”, describes a unique method to collect exosomes from single cells. The idea here is to plate the cells through a  mesh with small holes, so statistically only one cell enters and rests at the bottom (which they verify microscopically). They then remove the mesh and place on top of the cells a glass slide coated with antibodoes against known exosomal markers. The glass is placed 100micron above the cells. They then image the cells while simultaneously collect exosomes over time.


How to collect exosomes from single cells. Source: Chiu YJ et al, (2016) Small ePub: 2 Jun 2016.

The exosomes bound to the glass are then labeled by quantum dots and the number of Qdots for each cell location is counted under the microscope. [The authors don’t mention if and how they verify that there is only one Qdot per exosomes. For clarity, lets assume also a 1:1 ratio].

They tested and show that the maximum capacity of  of the exosome-Qdot complex in the 100x field of view of their microscope was ~1500 Qdots. To stay in range, they collected exosomes  for 3 hrs, so that the highest measured level was ~500 exosomes.  They measured the background signal (of empty places in the mesh) was about 3% of the signal, probably due to exosome diffusion side-ways.

And here comes the interesting part: they measured exosome secretion rate of three cell lines. Two lines (MCF7 & MDA-MB-231) showed exosomes secretion rate of ~60-65 exosomes per cell per hour. The third line (MCF10A) was more prolific, secreting ~170 exosomes per cell per hour.

In the previous post, I calculated that the researchers added ~100,000 exosomes per cell to the recipient cells. If the numbers from this paper here are correct, it is the equivalent to the number of exosomes secreted by 588 MCF10A cells for an hour.

To me, this says that those assays that use isolated concentrated exosomes with ratios of 1000s to 100s of thousands of exosomes/cell, are really way too high compared to what might actually happen.

Even if the paper here is under-estimating the number of exosomes/cell/hour – how much underestimation are we talking about? 2-fold? 10-fold? The assays described in the preious post used ~500-fold more exosomes that anticipated by the results here.

Of course, these are differnt cell lines (HEK293 were the cells used in the other paper), but how much differnc can there be? This needs to be tested.

A detailed protocol can be found at Bio-protocol, here.

Anyway, to me, this shows that the exosome field MUST obtain more quantitative data. Chiu YJ, Cai W, Shih YV, Lian I, & Lo YH (2016). A Single-Cell Assay for Time Lapse Studies of Exosome Secretion and Cell Behaviors. Small (Weinheim an der Bergstrasse, Germany) PMID: 27254278


5 responses to “Counting exosome secretion

  1. Pingback: The wild ride of the exosomes | greenfluorescentblog

  2. I believe there is more variability between cell lines than you might think. In my hands using NTA quantification I calculate on the order of 12,000 exosomes per MDA cell per hour. Almost 200-fold more than this paper describes, and not far off from the 500-fold ballpark for the mentioned assay. The approach the authors used in the cited paper assumes collecting the “common markers” accounts for an entire population of exosomes which, given the heterogeneity of exosome populations, may not be a valid assumption. I also doubt they are getting near 100% capture with this approach. Regardless of that point there is a recent paper comparing relative exosome secretion between different cell lines (Mesenchymal stem cell: An efficient mass producer of exosomes for drug delivery☆ Ronne Wee Yeh Yeo a,b,1, Ruenn Chai Lai a,1, Bin Zhang a, Soon Sim Tan a, Yijun Yin a, Bao Ju Teh a, Sai Kiang Lim). They show a 10,000 fold difference in secretion between high exosome-producing stem cells versus low exosome-producing myoblasts. In their hands HEK 293 cells released 5% the exosomes released by stem cells, which matches what I observe (typically around 6%). Quantitatively I have calculated release rates of 3,900 exosomes per hour for a single stem cell and 248 exosomes per hour for a single HEK cell.

    Liked by 1 person

    • Hi Scott,
      thanks for your input.
      I agree that by using a “common marker” the authors miss other types of exosomes and that in any case it is not 100% capture.
      I was unfamiliar with the data you cited and the numbers that you mentioned here. Where these number published?
      As to the data in figure 2 of this paper ( these are relative, not absolute numbers, and is over a 3-day period of an already 80% confluent culture. Was this normalized on the number of cells? (e.g. 80% confluence of HEKs or MEFs yields different cell numbers). What happens during those 3 days? do exosomes just remain in the media, inert? Or do some cell types re-absorb their exosomes more that others? Do all exosomes have similar stability in media over 3-day period? Did the culture reach an exosome “saturation” in the media at some point over these 3 days?
      It’s nice to have quantitative data, but i think the conclusions about the exosome production efficiency for each cell type might be over or under-rated.

      My feeling is that there is still lack of quantitative data on exosomes biology – numbers produced, numbers internalized by acceptor cells/by self, number of cargo molecules/single EV etc…
      There have been some publications, including those I wrote about here and some others. But on the whole, I think there is an immediate need for really hardcore quantitative analysis of these and other parameters.


      • In the paper I cited I do not think they normalize for cell count as dedicatedly as I have tried to do in some of my experiments (still unpublished). I count before seeding the flask, seed so that they can grow in the log phase over three days, and then count number of cells after three days to be able to use the “average number” of cells when I normalize. To your point about self-uptake of exosomes this is something that I think is really under-explored, and agree with you totally that more data is needed to answer this question. (At one point I was actually trying to design an experiment to look at that but it was quite the technical challenge). My only counter point would be secretion as I would describe it would probably be more aptly called “net secretion” or “effective secretion” to circumvent this phenomena, since likely an effective secretion is more relevant for most researchers’ experiments. I have come across a number of reports that suggest exosome accumulation in media is linear, which would indicate some interesting dynamics are at play (since If it was as simple as a per cell net secretion you’d expect accumulation to correlate perfectly with cellular growth curves – which are not linear). I’d also assume stability is. It the same for the exosomes of different cell lines given the diversity of lipid composition but haven’t seen a detailed paper on that yet either- so I’ll keep my eyes open to some more answers to these questions!

        Liked by 1 person

        • But why do you count after 3 days? why not after 3hrs? 12 hrs?
          3 days seems like a long time in my opinion to measure the outcome of what seems to be fairly quick events (1000’s of exosomes/hr means dozens of events per minute!). Is it just sensitivity of detection?
          As for self-uptake, here’s an idea for an experiment:
          split your culture to two sub-cultures.
          Collect exosomes from one culture.
          Count them, then fill them with a fluorescent dye.
          Apply fluorescently-labeled exosomes onto the other culture.
          Incubate for X time.
          Collect exosomes again.
          Count total exosomes and fluorescent exosomes.
          Subtract Fluorescent exosomes from culture 2 from total exosomes of culture 1.
          Normalize to cell number in culture 2.
          –> average number self-uptake exosomes/cell*time.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s