The development of life-threatening cancer metastases at distant organs requires disseminated tumour cells’ adaptation to, and co-evolution with, the drastically different microenvironments of metastatic sites1. Cancer cells of common origin manifest distinct gene expression patterns after metastasizing to different organs2. Clearly, the dynamic interaction between metastatic tumour cells and extrinsic signals at individual metastatic organ sites critically effects the subsequent metastatic outgrowth3,4. Yet, it is unclear when and how disseminated tumour cells acquire the essential traits from the microenvironment of metastatic organs that prime their subsequent outgrowth. Here we show that both human and mouse tumour cells with normal expression of PTEN, an important tumour suppressor, lose PTEN expression after dissemination to the brain, but not to other organs. The PTEN level in PTEN-loss brain metastatic tumour cells is restored after leaving the brain microenvironment. This brain microenvironment-dependent, reversible PTEN messenger RNA and protein downregulation is epigenetically regulated by microRNAs from brain astrocytes. Mechanistically, astrocyte-derived exosomes mediate an intercellular transfer of PTEN-targeting microRNAs to metastatic tumour cells, while astrocyte-specific depletion of PTEN-targeting microRNAs or blockade of astrocyte exosome secretion rescues the PTEN loss and suppresses brain metastasis in vivo. Furthermore, this adaptive PTEN loss in brain metastatic tumour cells leads to an increased secretion of the chemokine CCL2, which recruits IBA1-expressing myeloid cells that reciprocally enhance the outgrowth of brain metastatic tumour cells via enhanced proliferation and reduced apoptosis. Our findings demonstrate a remarkable plasticity of PTEN expression in metastatic tumour cells in response to different organ microenvironments, underpinning an essential role of co-evolution between the metastatic cells and their microenvironment during the adaptive metastatic outgrowth. Our findings signify the dynamic and reciprocal cross-talk between tumour cells and the metastatic niche; importantly, they provide new opportunities for effective anti-metastasis therapies, especially of consequence for brain metastasis patients.
Subscribe to Journal
Get full journal access for 1 year
only $3.90 per issue
All prices are NET prices.
VAT will be added later in the checkout.
Tax calculation will be finalised during checkout.
Rent or Buy article
Get time limited or full article access on ReadCube.
All prices are NET prices.
Quail, D. F. & Joyce, J.A. Microenvironmental regulation of tumor progression and metastasis. Nature Med. 19, 1423–1437 (2013)
Park, E. S. et al. Cross-species hybridization of microarrays for studying tumor transcriptome of brain metastasis. Proc. Natl Acad. Sci. USA 108, 17456–17461 (2011)
Joyce, J. A. & Pollard, J. W. Microenvironmental regulation of metastasis. Nature Rev. Cancer 9, 239–252 (2009)
Vanharanta, S. & Massagué, J. Origins of metastatic traits. Cancer Cell 24, 410–421 (2013)
Gray, J. Cancer: genomics of metastasis. Nature 464, 989–990 (2010)
Friedl, P. & Alexander, S. Cancer invasion and the microenvironment: plasticity and reciprocity. Cell 147, 992–1009 (2011)
Zhang, S. et al. Combating trastuzumab resistance by targeting SRC, a common node downstream of multiple resistance pathways. Nature Med. 17, 461–469 (2011)
Gonzalez-Angulo, A. M. et al. PI3K pathway mutations and PTEN levels in primary and metastatic breast cancer. Mol. Cancer Ther. 10, 1093–1101 (2011)
Wikman, H. et al. Relevance of PTEN loss in brain metastasis formation of breast cancer patients. Breast Cancer Res. 14, R49 (2012)
Song, M. S., Salmena, L. & Pandolfi, P. P. The functions and regulation of the PTEN tumour suppressor. Nature Rev. Mol. Cell Biol. 13, 283–296 (2012)
Hopkins, B. D. et al. A secreted PTEN phosphatase that enters cells to alter signaling and survival. Science 341, 399–402 (2013)
Miething, C. et al. PTEN action in leukaemia dictated by the tissue microenvironment. Nature 510, 402–406 (2014)
Mecha, M. et al. An easy and fast way to obtain a high number of glial cells from rat cerebral tissue: A beginners approach. Protocol Exchange http://dx.doi.org/10.1038/protex.2011.218 (2011)
Zhu, H., Han, C., Lu, D. & Wu, T. miR-17–92 cluster promotes cholangiocarcinoma growth: evidence for PTEN as downstream target and IL-6/Stat3 as upstream activator. Am. J. Pathol. 184, 2828–2839 (2014)
Liu, S.-Q., Jiang, S., Li, C., Zhang, B. & Li, Q.-J. miR-17–92 cluster targets phosphatase and tensin homology and Ikaros Family Zinc Finger 4 to promote TH17-mediated inflammation. J. Biol. Chem. 289, 12446–12456 (2014)
Olive, V. et al. miR-19 is a key oncogenic component of mir-17–92. Genes Dev. 23, 2839–2849 (2009)
Olive, V., Jiang, I. & He, L. mir-17-92, a cluster of miRNAs in the midst of the cancer network. Int. J. Biochem. Cell Biol. 42, 1348–1354 (2010)
Ventura, A. et al. Targeted deletion reveals essential and overlapping functions of the miR-17 through 92 family of miRNA clusters. Cell 132, 875–886 (2008)
Mittelbrunn, M. et al. Unidirectional transfer of microRNA-loaded exosomes from T cells to antigen-presenting cells. Nature Commun. 2, 282 (2011)
Suetsugu, A. et al. Imaging exosome transfer from breast cancer cells to stroma at metastatic sites in orthotopic nude-mouse models. Adv. Drug Deliv. Rev. 65, 383–390 (2013)
Frühbeis, C., Fröhlich, D. & Krämer-Albers, E. M. Emerging roles of exosomes in neuron–glia communication. Front. Physiol. 3, 119 (2012)
Kesimer, M. et al. Characterization of exosome-like vesicles released from human tracheobronchial ciliated epithelium: a possible role in innate defense. FASEB J. 23, 1858–1868 (2009)
Peinado, H. et al. Melanoma exosomes educate bone marrow progenitor cells toward a pro-metastatic phenotype through MET. Nature Med. 18, 883–891 (2012)
Conti, I. & Rollins, B. J. CCL2 (monocyte chemoattractant protein-1) and cancer. Semin. Cancer Biol. 14, 149–154 (2004)
Qian, B.-Z. et al. CCL2 recruits inflammatory monocytes to facilitate breast-tumour metastasis. Nature 475, 222–225 (2011)
Bonapace, L. et al. Cessation of CCL2 inhibition accelerates breast cancer metastasis by promoting angiogenesis. Nature 515, 130–133 (2014)
Lu, J. et al. 14-3-3ζ Cooperates with ErbB2 to promote ductal carcinoma in situ progression to invasive breast cancer by inducing epithelial-mesenchymal transition. Cancer Cell 16, 195–207 (2009)
Nagata, Y. et al. PTEN activation contributes to tumor inhibition by trastuzumab, and loss of PTEN predicts trastuzumab resistance in patients. Cancer Cell 6, 117–127 (2004)
Zhang, S. et al. Src family kinases as novel therapeutic targets to treat breast cancer brain metastases. Cancer Res. 73, 5764–5774 (2013)
Gril, B. et al. Effect of lapatinib on the outgrowth of metastatic breast cancer cells to the brain. J. Natl Cancer Inst. 100, 1092–1103 (2008)
We thank M.-C. Hung, H.-K. Lin and Z. Lu for reading the manuscript, A. Yung for PTEN promoter constructs, MD Anderson Cancer Center (MDACC) shRNA and ORFome, FACS, histology and high-resolution electron microscopy support, the animal core facilities (NIH CA16672) for technical support, and members of the Yu laboratory for helpful discussions. Thanks to A. Matsika for histological review and tissue microarray construction. This work was supported partially by DOD Center of Excellence grant (P.S.S.) subproject W81XWH-06-2-0033 (D.Y.), NIH Pathway to Independence Award 5R00CA158066-05 (S.Z.), DOD Postdoctoral Fellowship W81XWH-11-1-0003 (C.Z.), Isaiah Fidler Fellowship in Cancer Metastasis (F.J.L.), PO1-CA099031 project 4 (D.Y.), RO1-CA112567-06 (D.Y.), R01CA184836 (D.Y.), Susan G. Komen Breast Cancer Foundation Promise Grant KG091020 (D.Y.), METAvivor Research Grant (D.Y.), Breast and Ovarian Cancers Moon Shot program, China Medical University Research Fund, and Sowell-Huggins Pre-doctoral Fellowship (L.Z.) and Professorship (D.Y.) in Cancer Research. D.Y. is the Hubert L. & Olive Stringer Distinguished Chair in Basic Science at the MDACC.
The authors declare no competing financial interests.
Extended data figures and tables
a, Schematics of microarray analyses. Patients’ brain metastases exhibited a discrete gene expression profile with 650 genes significantly downregulated compared to bone or lung metastases (GSE14020). Cancer cells were injected into immunodeficient mice to produce orthotopic primary tumours (MDA-MB-231 cells for mammary tumour, PC14 for prostate tumour, A375SM for melanoma) and experimental brain metastases (all three lines). Brain metastases derived from these three cancer cell lines exhibited 2,161 commonly downregulated genes compared to their respective primary tumours (GSE19184). PTEN is one of only 54 commonly downregulated genes in brain metastases of both data sets. b, Heat-maps showing expression of 54 commonly downregulated genes (see a) in clinical brain metastases versus lung metastases and bone metastases. c, Heat-maps showing expression of the 54 genes (see a) in cell-line-induced primary tumours versus experimental brain metastases. d, Kaplan–Meier survival analyses showing no significant differences in brain metastasis-free survival between breast cancer patients with primary tumours expressing normal PTEN or low PTEN mRNA in GEO cDNA microarray set GSE2603 (P = 0.74). e, PTEN mRNA levels detected in primary breast tumours from patients with or without brain metastasis relapse. Three GEO cDNA microarray data sets (GSE2034, GSE2603 and GSE12276) with clinical annotation were analysed. Relative PTEN expression levels were compared by t-test.
Extended Data Figure 2 PTEN expression in different metastatic organ microenvironments and in vitro culture condition.
a, Breast cancer cell lines (MDA-MB-231, HCC1954, BT474, and MDA-MB-435) were cultured and injected either to the MFP to form primary tumour or intracarotidly to form brain metastases. Cell pellets and tumour tissues were stained for PTEN expression using anti-PTEN antibodies as described previously28. b, IHC staining of PTEN in brain metastases, paired lung metastases and primary tumour derived from either MDA-MB-231 or 4T1 cells. PTEN expression level was analysed based on an IRS scoring system. c, PTEN mRNA levels between parental MDA-MB-231 and CN-34 breast cancer cell lines (blue) and their brain-seeking sublines (red). Normalized PTEN-specific probe intensity values were extracted from cDNA microarray data set GSE12237. Dot plot shows the mean probe intensity derived from independent RNA samples. d, PTEN qRT–PCR (mean ± s.e.m., t-test) and PTEN IHC in MDA-MB-231Br secondary tumours and cultured cells (3 biological replicates, with 3 technical replicates each). e, f, qRT–PCR (e) and western blot (f) analysis of PTEN mRNA expression (mean ± s.e.m., t-test) or protein expression in MDA-MB-231 cells after co-culture with either primary mouse CAFs isolated from MDA-MB-231 xenograft tumours or primary mouse glia isolated from mouse brain (3 biological replicates, with 3 technical replicates each). g, Representative methylation-specific PCR of PTEN promoter and quantification under co-culture with glia or CAF (mean ± s.e.m., t-test, 2 biological replicates, with 2 technical replicates each). h, PTEN promoter activity measured by luciferase reporter in HCC1954 cells after co-culture with either CAF or glia cells for 48 h (mean ± s.e.m., t-test, P = 0.5271, 3 biological replicates, with 3 technical replicates each).
a, IHC analyses of the expression of AMP-activated protein kinase (AMPK), pro-apoptotic protein BIM and transcription factor E2F1 (mean ± s.d., t-test) in brain metastasis tumours with/without pre-knocking out the miR-17˜92 cluster in the brain microenvironment. B, brain tissue; M, brain metastases. b, Schematic of experimental design. The Ad-GFAP-Cre adenovirus was injected intracranially to the right hemisphere of the Mirc1tm1.1Tyj/J mouse, and the control adenovirus (Ad-βGLuc) was injected intracranially to contralateral side of the brain. B16BL6 cells were then injected intracranially to both sides. c, IHC analysis of Cre expression in the brain astrocytes. d, IHC analysis of PTEN expression in the tumour cells. e, Quantification of PTEN expression in tumour cells (mean ± s.d., t-test). f, Quantification of intracranial tumour outgrowth by volume (mean ± s.e.m., t-test). g, qRT–PCR analyses of miR-19a and PTEN mRNA in tumour cell HCC1954 after 48 h co-culture with primary astrocytes from Mirc1tm1.1Tyj/J mice pre-infected (48 h) with adenovirus (Ad-βGLuc or Ad-GFP-Cre) (mean ± s.e.m., t-test, 3 biological replicates, with 3 technical replicates each). h, Western blot of PTEN protein in the indicated tumour cells co-cultured as in g. i, Knockdown of miR-17˜92 allele in cultured primary astrocytes. miR-17˜92 cluster is flanked by loxP site in Mirc1tm1.1Tyj/J mouse. Primary astrocytes were isolated from Mirc1tm1.1Tyj/J mouse brain then infected by adenovirus encoding for βGLuc or GFP-Cre protein. Concentrated adenovirus particles of indicated volume (same MOI ˜108 U ml−1) encoding βGLuc or GFP-Cre proteins were added to 106 astrocytes. Left, representative image showing the infection efficiency. Right, bar diagram showing the relative miR-19a expression (one of the five miRNA genes in the miR-17~92 cluster) three days after adenovirus infection (mean ± s.e.m., t-test, 3 biological replicates, with 3 technical replicates each).
Extended Data Figure 4 Contact-independent downregulation of PTEN in tumour cells by miR-19a from astrocyte-derived exosomes.
a, Flow cytometric detection of Cy3-miR-19a and FITC-EpCAM in tumour cells 60 h after co-culture with Cy3-miR-19a-transfected astrocytes and CAFs. b, c, Tumour cells were co-cultured with conditioned media from astrocytes or CAFs for 60 h. RT–PCR analyses of the PTEN-targeting miR-19a level (b) and PTEN mRNA level (c) in tumour cells (mean ± s.e.m., t-test, 3 biological replicates, with 3 technical replicates each). d, Western blot detecting PTEN protein levels in HCC1954 cells after culture with conditioned media from either astrocytes or CAFs for 60 h. e, Flow cytometry detecting CD63+ exosomes extracted from CAF- or astrocyte-conditioned media. f, Histogram showing the exosome protein level detected from CAF- and astrocyte-conditioned media normalized by cell number (mean ± s.e.m., t-test, P < 0.0001, 3 biological replicates, with 3 technical replicates each). g, RT–PCR analyses of miR-19a level in exosomes extracted from CAF- or astrocyte- conditioned media normalized by equal cell numbers (mean ± s.e.m., t-test, P < 0.0001, 3 biological replicates, with 3 technical replicates each).
a, Exosome-releasing inhibitor (DMA) treatment reduced exosome secretion from astrocytes compared to vehicle treated astrocytes. Astrocytes were treated with DMA (25 μg ml−1) or vehicle for 4 h; exosomes were concentrated from astrocyte-conditioned media and total proteins from exosomes were examined by BCA assay (normalized to total cell numbers) (mean ± s.e.m., t-test, P = 0.038, 3 biological replicates, with 3 technical replicates each). b, Knockdown of Rab27a in astrocytes by siRNA. Two siRNAs targeting mouse Rab27a were transiently transfected into astrocytes, and the Rab27a mRNA level was examined by RT–PCR 48 h after transfection (mean ± s.e.m., t-test, P < 0.01, 3 biological replicates, with 3 technical replicates each). c, Knocking down Rab27a in astrocytes inhibited exosome release. Forty-eight hours after Rab27a-targeting siRNAs were transfected, exosomes were collected from astrocyte-conditioned media and total proteins from exosomes were examined by BCA assay (normalized to total cell numbers) (mean ± s.e.m., t-test, 3 biological replicates, with 3 technical replicates each). d, Histogram showing relevant changes of Rab27a and Rab27b mRNA level in primary astrocytes infected with pLKO.shRab27a or pLKO.shRab27b virus (mean ± s.e.m., t-test, P < 0.001, 3 biological replicates, with 3 technical replicates each). e, Change of exosome protein level detected in the conditioned media from astrocytes infected by pLKO.shRab27a or pLKO.shRab27b virus by BCA assay (normalized to total cell numbers) (mean ± s.e.m., t-test, P < 0.001, 3 biological replicates, with 3 technical replicates each). f, g, IHC analysis showing the expression level of Rab27a and Rab27b (f) and exosome marker expression CD63 (g) in the brain tissue derived from mice injected with control lentivirus or Rab27a/b shRNA lentiviruses and subsequently intracranially injected with B16BL6 cells. h, i, IHC analysis showing the expression level of Rab27a and Rab27b (h) and exosome marker expression CD63 (i) in the brain tissue derived from mice injected with Rab27a/b shRNA lentiviruses and subsequently intracranially injected with B16BL6 cells and vehicle at the left side or B16BL6 cells and astrocyte-derived exosomes at the right side.
Extended Data Figure 6 Brain extravasation of MDA-MB-231 parental cells with or without induction of doxycycline-inducible PTEN shRNA knockdown, PTEN expression and CCL2 shRNA knockdown.
a, Western blot showing PTEN expression levels after treating MDA-MD-231 cells with doxycycline. MDA-MD-231 cells were stably infected with inducible shRNA expression vectors (pTRIPZ-control-shGFP as control and pTRIPZ-shRNA-RFP for PTEN shRNA). Doxycycline (1 μg ml−1) was added to induce shRNA expression for 5 days. As indicated, doxycycline was withdrawn in some samples for another 5 days before analysis. b, Schematics of in vivo extravasation assay. shControl-GFP and shPTEN-RFP cells were mixed at a 1:1 ratio. In total, 200,000 cells were ICA injected into mice, and doxycycline (50 μg kg−1) was given intraperitoneally daily. Brains were collected 5 days after ICA injection. c, Dot plot of extravasated cell counts 5 days after ICA injection of indicated MDA-MB-231 sublines. Tumour-bearing brains were collected and sectioned into 100 μm coronal slices. Extravasated tumour cells were counted under the fluorescence microscope (mean ± s.d., t-test). d, MDA-MB-231Br single cells were expanded into subclones (C12, C14, C18 and C19), which were transfected with doxycycline-inducible pTRIPZ-RFP or pTRIPZ-PTEN. 48 h after doxycycline (1 μg ml−1) treatment, PTEN induction was tested by western blotting. The C14 clone was used for further in vivo assays (see e, f and Fig. 4a). e, IHC staining of induced PTEN expression in brain metastases derived from mice injected with MDA-MB-231Br (231Br-RFP or 231Br-PTEN) cells. f, IHC analysis of PTEN downstream signalling pathway, including phosphorylated pAkt(T308), pAkt(S473) and pP70S6K(T389+T412) in brain metastases from mice injected with 231Br-RFP or 231Br-PTEN cells. Top, dot plot of IHC data quantification by IRS (mean ± s.d., t-test); bottom, representative IHC staining data. g, Histograms of PTEN and CCL2 mRNA levels (mean ± s.e.m., t-test) in indicated cancer cell lines 48 h after transfection with control or PTEN siRNAs (3 biological replicates, with 3 technical replicates each). h, Histogram showing the inducible CCL2 knockdown. MDA-MB-231Br cells were stably infected with pTRIPZ-inducible CCL2 shRNAs. 48 h after doxycycline (1 μg ml−1) treatment, CCL2 mRNA was examined by RT–PCR (mean ± s.e.m., t-test, 3 biological replicates, with 3 technical replicates each). i, Doxycycline-induced CCL2 knockdown in brain metastases. Mice were ICA injected with MDA-MB-231Br cells containing control or CCL2 shRNAs. Doxycycline (50 μg kg−1) was given to mice intraperitoneally daily after injection. IHC staining of CCL2 expression levels in brain metastases derived from MDA-MB-231Br cells. T, brain metastasis tumours at day 30 after ICA injection.
a, Heat-map showing differentially expressed protein markers of reverse-phase protein array analysis. MDA-MB-231Br cells were stably infected with pTRIPZ-RFP or pTRIPZ-PTEN (231Br-RFP or 231Br-PTEN) and induced by doxycycline (1 μg ml−1) for 48 h. b, Box chart showing the absolute intensity of PTEN and NF-κB p65(S536). c, Western blot analysis of NF-κB p65 nuclear translocation, after cells were treated with Akt inhibitor MK2206 (10 μg ml−1) 24 h before separation into cytosolic (Cyto) and nuclear (Nuc) fractions. d, Western blot analysis of NF-κB p65 nuclear translocation, after cells were treated with NF-κB inhibitor PDTC (0.2 mM) 16 h before separation into cytosolic and nuclear fractions. e, Relative CCL2 mRNA expression after NF-κB inhibitor PDTC treatment analysed by qRT–PCR (mean ± s.e.m., t-test, 3 biological replicates, with 3 technical replicates each). Cells were treated with PDTC (0.2 mM) for 16 h. f, Relative CCL2 protein expression after PDTC treatment analysed by ELISA (mean ± s.e.m., t-test, 3 biological replicates, with 3 technical replicates each). Cells were treated as in e.
a, Co-expression of IBA1 and CCR2 on myeloid cells freshly isolated from mouse brain by CD11b beads. Representative immunofluorescence staining of IBA1 (left). FACS analysis of CD11b+ cells for CCR2 expression. b, Relative CCR2 expression in the BV2 microglia cell line compared with NIH3T3 fibroblasts. CCR2 mRNA level analysed by qRT–PCR (mean ± s.e.m., t-test, 3 biological replicates, with 3 technical replicates each) (left) and protein expression analysed by FACS (right) c, Transwell migration assay examining the directional migration of BV2 cells towards CCL2. In total, 105 BV2 cells were seeded in the top chamber of the transwell units, and CCL2 or BSA (20 ng ml−1) was added into serum-free media in the bottom chamber. The migrated cell numbers were counted at 24 h. Next, CCR2 antagonists with different concentrations (10 μM, 1 μM and 0.1 μM) were added into the top chamber with BV2 cells, and CCL2 (20 ng ml−1) was added into serum-free media in the bottom chamber. The migrated cell numbers were counted at 24 h. d, Quantification of BV2 cell migration assay (mean ± s.e.m., t-test, 3 biological replicates, with 3 technical replicates each).
Extended Data Figure 9 The association between PTEN, CCL2 expression and recruitment of IBA1+ myeloid cells in patients’ brain metastases and matched primary breast tumours.
a, Summary histogram of CCL2 protein levels in primary breast tumours and matched brain metastases from 35 patients. Chi-square test was used to compare the IHC score in primary breast tumours versus matched brain metastases. P < 0.05 is defined as significantly different. b, Tables showing IHC scores of PTEN and CCL2 expression in primary breast tumours and matched brain metastases. c, Representative IHC staining of CCL2 proteins and IBA1+ myeloid cells in patients’ brain metastases, and the correlation plot showing the Pearson correlation between CCL2 and IBA1 staining in patients’ brain metastases (R = 0.371, P = 0.028).
Extended Data Figure 10 PTEN loss induced by astrocyte-derived exosomal microRNA primes brain metastasis outgrowth via functional cross-talk between disseminated tumour cells and brain metastatic microenvironment.
Top, disseminated tumour cells extravasate into the brain. a–-c, Exosomes secreted by astrocytes in the brain microenvironment transfer PTEN-targeting miRNA into extravasated brain metastatic tumour cells, leading to PTEN downregulation in tumour cells. c, d, PTEN loss in brain metastatic tumour cells increases their CCL2 secretion, facilitating the recruitment of IBA1+/CCR2+ myeloid cells at the micrometastasis site. d, e, The recruited IBA1+ myeloid cells enhance proliferation and inhibit apoptosis of metastatic tumour cells, and promote metastatic outgrowth.
About this article
Cite this article
Zhang, L., Zhang, S., Yao, J. et al. Microenvironment-induced PTEN loss by exosomal microRNA primes brain metastasis outgrowth. Nature 527, 100–104 (2015). https://doi.org/10.1038/nature15376
Extracellular vesicle-derived miRNA as a novel regulatory system for bi-directional communication in gut-brain-microbiota axis
Journal of Translational Medicine (2021)
LncRNA SPOCD1-AS from ovarian cancer extracellular vesicles remodels mesothelial cells to promote peritoneal metastasis via interacting with G3BP1
Journal of Experimental & Clinical Cancer Research (2021)
Molecular Biomedicine (2021)
Co-continuous structural effect of size-controlled macro-porous glass membrane on extracellular vesicle collection for the analysis of miRNA
Scientific Reports (2021)
British Journal of Cancer (2021)