Diagnosis by extracellular vesicles

The detection of a single molecule anchored to circulating extracellular vesicles allows late-stage pancreatic cancer to be identified from just one drop of a patient's blood. See Article p.177

On page 177 of this issue, Melo et al.1 describe a non-invasive test that identifies patients with late-stage pancreatic cancer with 100% certainty, and that can distinguish patients with precancerous pancreatic lesions from those with benign pancreatic diseases. Although the number of patients in the precancerous-lesion group was low, and the findings require further validation in a larger cohort, the potential implications of such a test are huge. It would allow clinicians to decide whether or not to perform potentially debilitating surgery.

The test involves detecting a membrane-anchored proteoglycan molecule, glypican-1 (GPC1), in vesicles that circulate in the bloodstream. The authors found this proteoglycan in membranous material isolated from a small amount of frozen serum taken from all tested patients who had pancreatic cancer. By contrast, the sera of patients with other pancreatic diseases did not contain higher levels of GPC1-containng (GPC1+) vesicles than those of healthy donors (Fig. 1). The test was more reliable than a commonly used assay (which involves the ELISA method) to detect the presence in whole blood of a pancreatic-tumour biomarker called carbohydrate antigen 19-9 (CA 19-9). About half of the patients without cancer had elevated CA 19-9 levels, whereas none had elevated GPC1+ vesicles, and CA 19-9 was not elevated above control levels in many patients with cancer. Furthermore, in a mouse model of genetically induced pancreatic cancer, Melo and colleagues' test gave positive results before a detectable tumour was present.

Figure 1: GPC1 distinguishes pancreatic cancer from benign disease.

Melo et al.1 show that extracellular vesicles isolated from the bloodstream of patients with precancerous pancreatic lesions or pancreatic cancer contain the membrane-anchored proteoglycan GPC1, whereas this molecule is not found at higher-than-normal levels in vesicles in the blood of patients with non-cancerous pancreatic disease or of healthy donors.

Overexpression of GPC1 in pancreatic carcinoma and a positive role for this overexpression in tumour proliferation and metastasis have been reported previously using tumour cell lines and mouse models2,3,4. The novelty of Melo and colleagues' report resides in the presence of GPC1 in circulating vesicles in serum, and in the striking value of this molecule as a biomarker. Note that simple detection of serum GPC1 by ELISA, without concentrating the vesicles, does not provide a more reliable diagnostic test than CA 19-9 detection. Thus, this work demonstrates for the first time that circulating vesicles in blood can be a source of specific and reliable diagnostic biomarkers for cancer.

Vesicles present in bodily fluids, which are collectively known as extracellular vesicles (EVs) or exosomes (as Melo and colleagues call them), have been investigated as potential biomarkers for various diseases for a decade5,6,7. But until now, overexpression of exosomes or exosomal markers was observed either only in advanced disease or after (rather than before) detectable tumour progression8, or without statistical significance9. Recently published exosome analyses of the blood of patients with lung10 or pancreatic cancer11 have reported cancer detection with 75% and 93% specificity, respectively. However, the studies respectively measured the expression of a combination of 30 proteins on a microarray chip or a combination of 5 proteins and 4 microRNA molecules. These tests are less reliable and more complex than Melo and colleagues' test, which involves the detection of a single molecule and more-conventional techniques.

The authors' protocol uses long ultracentrifugation of small volumes of serum, coating of beads with the resulting pellet, and staining of the beads with a GPC1-specific antibody before analysing them by flow cytometry. Ultracentrifuges and flow cytometers are widespread and straightforward to use, suggesting that this protocol could be implemented in clinical laboratories as a routine procedure for evaluating patients who present with symptoms of pancreatic disease. The authors also show that the beads that have captured GPC1+ vesicles contain a mutant messenger RNA expressed by the tumour, which could allow further exploration of tumour properties.

On a slightly disappointing note, it seems that this test might not be useful for cancers other than pancreatic cancer. Although the authors' identification of GPC1 as a cancer-specific protein secreted in exosomes involved comparing cancerous and non-cancerous cell lines of breast origin, expression of GPC1 in circulating vesicles did not reliably identify patients with breast cancer, nor allow patients to be assigned to a specific breast-cancer subtype. However, GPC1 expression in breast cancer may still warrant further exploration. I noticed that among those patients whose blood EVs showed GPC1 expression, there were two distinct populations, with either a high or an intermediate number of GPC1+ EVs. The authors did not discuss this observation, but I wonder if the amount of GPC1+ circulating EVs could provide additional diagnostic or prognostic information.

Finally, I would like to devote a few words to the term exosome. It was first used in the context of vesicles in 1981 to describe membrane-enclosed structures of variable size (either 40 nanometres or 500–1,000 nm in diameter) that had been 'exfoliated' from the surface of cultured cells12. The same term was then proposed in 1987 for small (50–100 nm) vesicles that form inside cellular compartments called endosomes, and that are released extracellularly when these compartments fuse with the cell membrane13. Several research groups, including mine, have defended this latter use, but as EVs have become the focus of increasing interest, the word exosome has started to be used for small EVs without showing that they arose from endosomes rather than from the cell membrane14.

The EVs used by Melo et al. for GPC1-based diagnostics are recovered from an untracentrifugation pellet that will contain exosomes as well as other types of small EVs, lipoproteins and even complexes of proteins and nucleic acids. Because GPC1 is a membrane-anchored protein, it is probably recovered in EVs, but the authors do not show the origin of the EVs that have diagnostic value. Melo and colleagues' paper will probably contribute to the increasing popularity of the term exosome, and possibly a generalization of its use to any type of small EV, about which purists such as myself can probably do little. That said, the intracellular origin of circulating GPC1 is irrelevant to its use in a diagnostic test, and a semantic issue should not interfere with the diffusion of such clinically important findings.Footnote 1


  1. 1.

    See all news & views


  1. 1

    Melo, S. A. et al. Nature 523, 177–182 (2015).

    ADS  CAS  Article  Google Scholar 

  2. 2

    Kleeff, J. et al. J. Clin. Invest. 102, 1662–1673 (1998).

    CAS  Article  Google Scholar 

  3. 3

    Aikawa, T. et al. J. Clin. Invest. 118, 89–99 (2008).

    CAS  Article  Google Scholar 

  4. 4

    Whipple, C. A., Young, A. L. & Korc, M. Oncogene 31, 2535–2544 (2012).

    CAS  Article  Google Scholar 

  5. 5

    Kim, J. W. et al. Clin. Cancer Res. 11, 1010–1020 (2005).

    ADS  CAS  PubMed  Google Scholar 

  6. 6

    Zhou, H. et al. Kidney Int. 70, 1847–1857 (2006).

    CAS  Article  Google Scholar 

  7. 7

    Skog, J. et al. Nature Cell Biol. 10, 1470–1476 (2008).

    CAS  Article  Google Scholar 

  8. 8

    Peinado, H. et al. Nature Med. 18, 883–891 (2012).

    CAS  Article  Google Scholar 

  9. 9

    Logozzi, M. et al. PLoS ONE 4, e5219 (2009).

    ADS  Article  Google Scholar 

  10. 10

    Jakobsen, K. R. et al. J. Extracell. Vesicles 4, 26659 (2015).

    Article  Google Scholar 

  11. 11

    Madhavan, B. et al. Int. J. Cancer 136, 2616–2627 (2015).

    CAS  Article  Google Scholar 

  12. 12

    Trams, E. G., Lauter, C. J., Salem, N. Jr & Heine, U. Biochim. Biophys. Acta 645, 63–70 (1981).

    CAS  Article  Google Scholar 

  13. 13

    Johnstone, R. M., Adam, M., Hammond, J. R., Orr, L. & Turbide, C. J. Biol. Chem. 262, 9412–9420 (1987).

    CAS  PubMed  Google Scholar 

  14. 14

    Gould, S. J. & Raposo, G. J. Extracell. Vesicles 2, 20389 (2013).

    Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Clotilde Théry.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Théry, C. Diagnosis by extracellular vesicles. Nature 523, 161–162 (2015).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.