Conceived and designed the experiments: AW NG UH GR. Performed the experiments: NG UH. Analyzed the data: AW NG UH GR. Contributed reagents/materials/analysis tools: AW NG UH GR. Wrote the paper: AW NG UH GR.
The authors have declared that no competing interests exist.
Shedding microvesicles are membrane released vesicles derived directly from the plasma membrane. Exosomes are released membrane vesicles of late endosomal origin that share structural and biochemical characteristics with prostasomes. Microvesicles/exosomes can mediate messages between cells and affect various cell-related processes in their target cells. We describe newly detected microvesicles/exosomes from cardiomyocytes and depict some of their biological functions.
Microvesicles/exosomes from media of cultured cardiomyocytes derived from adult mouse heart were isolated by differential centrifugation including preparative ultracentrifugation and identified by transmission electron microscopy and flow cytometry. They were surrounded by a bilayered membrane and flow cytometry revealed presence of both caveolin-3 and flotillin-1 while clathrin and annexin-2 were not detected. Microvesicle/exosome mRNA was identified and out of 1520 detected mRNA, 423 could be directly connected in a biological network. Furthermore, by a specific technique involving TDT polymerase, 343 different chromosomal DNA sequences were identified in the microvesicles/exosomes. Microvesicle/exosomal DNA transfer was possible into target fibroblasts, where exosomes stained for DNA were seen in the fibroblast cytosol and even in the nuclei. The gene expression was affected in fibroblasts transfected by microvesicles/exosomes and among 333 gene expression changes there were 175 upregulations and 158 downregulations compared with controls.
Our study suggests that microvesicles/exosomes released from cardiomyocytes, where we propose that exosomes derived from cardiomyocytes could be denoted “cardiosomes”, can be involved in a metabolic course of events in target cells by facilitating an array of metabolism-related processes including gene expression changes.
Many cells have the capacity to release microvesicles of endocytic origin to the extracellular space. The first cellular system to be explored in this regard was the prostate acinar cells. We showed more than 30 years ago that human prostatic fluid and therewith seminal plasma contains membrane surrounded, nano-sized (40–490 nm), secretory granules and microvesicles
Exosomes constitute the correspondence to prostasomes in other cell types. Exosome is nowadays used as a family name where prostasome exemplifies a specific name denoting the cell type of origin. Exosomes were originally thought to provide unconventional means for removal of redundant membrane proteins from reticulocytes
Since prostasomes can interact with spermatozoa
We now demonstrate that the cardiomyocyte, generally not considered secretory, releases microvesicles/exosomes with capacity to transfer genetic information to target cells.
HL-1, a cardiomyocyte cell line derived from adult mouse heart
During culture, the medium was changed routinely every 24 h.
NIH 3T3 cells
All cultures were kept in an atmosphere of 95% air-5% CO2, 37°C at a relative humidity of approximately 95%.
After 72 h, cells were grown to confluence and all media were replaced with serum-free and antibiotic-free media for 24 or 48 h. Media were collected and the cells were then washed with ice-cold phosphate buffered saline (PBS) (without calcium and magnesium, Fisher Scientific), and subsequently harvested by scraping and placed in RNAlater (Qiagen).
Media were centrifuged to remove cell debris, 3,000×g for 20 min at 4°C, repeated three times, followed by 10,000×g for 20 min at 4°C, repeated three times. The acquired supernatants were ultracentrifuged at 130,000×g (49,000 rpm) for 2 h at 4°C in an MLS-50 rotor and a Beckman Optima™ MAX-E Ultracentrifuge (Beckman Coulter) to separate microvesicle/exosome pellet and medium supernatant containing soluble molecules. Pellet was dissolved in PBS. A similar ultracentrifugation of the fetal bovine serum alone was carried out to rule out the presence of any microvesicles/exosomes in the culture medium.
Fluorescence-activated cell sorter (FACS) was used to detect proteins on microvesicle/exosome surfaces. Isolated microvesicles/exosomes were stained with 250 ng mouse anti-annexin-2, mouse anti-clathrin heavy chain, mouse anti-flotillin-1 and mouse anti-caveolin-3 (BD Biosciences) in 100 µL PBS, for 20 min, in the dark on ice. After adding 1.9 mL PBS and an additional ultracentrifugation to wash the pellet, it was resuspended in 100 µL PBS and 1 µL rat anti-mouse IgG phycoerythrin (PE) and incubated for 20 min, in the dark on ice. The ultracentrifugation was repeated and the pellet resuspended in PBS. Microvesicles/exosomes were analyzed on FACSCalibur (Becton Dickinson).
To evaluate if DNA is present inside or outside the microvesicles/exosomes, the FACS was used for analysis on DNA stained microvesicles/exosomes. 250 µL of microvesicles/exosomes (from a 1 mL microvesicle/exosome stock solution) was incubated with acridine orange (AO) (Invitrogen) to a final concentration of 20 µmol/L. AO is membrane permeable under ordinary conditions according to the manufacturer. Another 250 µL of the microvesicle/exosome stock solution was incubated with propidium iodide (PI) (Sigma) to a final PI concentration of 30 µmol/L. As a positive control an additional 250 µL of the microvesicle/exosome stock solution was incubated with PI and genomic DNA to a final PI concentration of 30 µmol/L. PI is not membrane permeable under ordinary conditions according to the manufacturer. As control, unstained microvesicles/exosomes, (from the stock solution) treated in the same way as the stained samples, was used. Prepared staining solutions (PI (30 µmol/L) and AO (20 µmol/L), respectively) were incubated for 90 min at 21°C, protected from light. The microvesicles/exosomes in the respective staining solution were pelleted by ultracentrifugation at 100,000×g for 2 h. Supernatants were discarded and pellets were washed twice with PBS and the pellets were resolved in PBS and adjusted to the same starting volume (1 mL).
The DNA-stained microvesicle/exosome samples were prepared for flow cytometric analysis by adding 500 µL of the samples to four tubes. The fluorescence was analyzed using a FACSCalibur (Becton Dickinson) at appropriate fluorescence emitting wavelengths; 525±15 nm for AO and 670/Long pass (LP) for PI.
For electron microscopy the microvesicles/exosomes were fixed in a solution containing 3% glutaraldehyde in 75 mmol/L sodium cacodylate buffer (pH 7.4) with 4% polyvinylpyrolidone and 2 mmol/L CaCl2, for 6 h. They were subsequently rinsed in the same buffer for one hour, and then post-fixed in 1% osmium tetroxide over night at 4°C. After another rinse in buffer the sample was dehydrated in a graded series of acetone and then embedded in an epoxy resin.
Ultrathin sections (70 nm) were cut, and collected on formvar coated copper grids and then contrasted with uranyl acetate and lead citrate for electron microscopy performed with a JEOL 1200-EX (Jeol Ltd.).
DNA was isolated with GenElute Mammalian Genomic DNA Miniprep Kit (Sigma-Aldrich) from a microvesicular pellet prepared from 18 mL Claycomb medium after 48 h incubation with cardiomyocytes. To add a poly-T tail, DNA was incubated with 25 µL 100 mmol/L dGTP (Gibco BRL, Life Technologies) and terminal deoxynucleotidyl transferase (TdT) (Invitrogen) for 30 min at 37°C, according to manufacturer's protocol. The constructed cDNA was purified and transcribed to synthesize biotinylated cRNA with Illumina Totalprep RNA Amplification Kit (Ambion). Total RNA was isolated with RNEASY Mini Kit (Qiagen) from cardiomyocytes and microvesicular/exosomal pellets prepared from 18 mL Claycomb medium after 48 h incubation with cardiomyocytes. Aliquots of RNA were converted to biotinylated double-stranded cRNA according to the specifications of the Illumina Totalprep RNA Amplification Kit (Ambion). The labeled cRNA samples were hybridized to MouseRef-8 Expression Beadchip (Illumina), incubated with streptavidin-Cy3 and scanned on the Illumina Beadstation GX. Illumina Beadstudio software, version 3.3.7. was used to normalize intensity data with the Beadstudio cubic spline algorithm. Significance detection P-value was set at <0.01. All data are MIAME compliant and are available through NCBIs Gene Expression Omnibus (GEO) and are accessible through GEO Series accession number GSE21707. To avoid detecting false positive genes due to low signal intensity, a minimum signal intensity of >50 was utilized (2.5 times highest background signal). Identified mRNAs had to be detected both in cardiomyocytes and microvesicles/exosomes to be considered as positively detected. Since cardiomyocytes reasonably are the source of microvesicular/exosomal mRNA, they should themselves contain the same mRNA.
Microvesicles/exosomes were stained with 20 µmol/L AO (Invitrogen) for 90 min, in the dark at room temperature. The sample was diluted to 4 mL and ultracentrifuged at 130,000×g (49,000 rpm) for 2 h at 4°C. The supernatant was removed to eliminate contamination of unincorporated AO. The AO-stained microvesicles/exosomes suspended in 1 mL PBS were then put in a dialysis bag with a 3,500 MWCO dialysis membrane (Spectra/Por) and dialysed against 300 mL PBS for 24 h with one change of dialysis buffer after 5 h. The sample was ultracentrifuged (130,000×g for 2 h at 4°C) and the pellet dissolved in DMEM and incubated for 3 h with fibroblasts, grown for 24 h on a cell culture microscope slide (Falcon). The slide was subsequently mounted with DAPI to stain fibroblast nuclei and studied in a Nikon Eclipse E800 confocal microscope. Light microscope was used to add a layer in images to visualize cell borders.
Fibroblasts, grown on 6-well plates, were incubated for 48 h with serum-free and antibiotic-free Claycomb medium, previously incubated for 24 h with cardiomyocytes.
A part of the same Claycomb medium was ultracentrifuged, 130,000×g for 2 h at 4°C and the supernatant was also incubated for 48 h with fibroblasts. Isolated microvesicles/exosomes from cardiomyocytes incubated for 48 h in Claycomb media were dissolved in DMEM and incubated with fibroblasts for 48 h. These stimulated fibroblasts were compared to control fibroblasts incubated in fresh Claycomb medium and DMEM, respectively. RNA was prepared from fibroblasts, labelled and hybridized to MouseRef-8 Expression Beadchip (Illumina), as described above.
To determine differentially expressed genes microarray data were analyzed using gene expression module in Beadstudio software, version 3.3.7. Intensity data were normalized using the Beadstudio cubic spline algorithm. Significant differential expression was calculated using the Illumina Beadstudio software by applying normalization using the Beadstudio cubic spline algorithm and multiple testing corrections using Benjamini and Hochberg False Discovery Rate (FDR)
Filtering step | Fb with Claycomb medium | Fb with Claycomb medium supernatant | Fb with ultra- centrifuged pellet | |||
DifferentialP-value<0.05 | 1384 | 501 | 656 | |||
695↑ | 689↓ | 201↑ | 300↓ | 280↑ | 376↓ | |
↓ | ↓ | ↓ | ||||
FDR | 400 | 102 | 249 | |||
213↑ | 187↓ | 21↑ | 81↓ | 88↑ | 161↓ | |
↓ | ↓ | ↓ | ||||
Detection | 335 | 96 | 209 | |||
P-value<0.05 | 177↑ | 158↓ | 21↑ | 75↓ | 75↑ | 134↓ |
↓ | ↓ | ↓ | ||||
Foldchange>1.5, <0.67 | 333 | 96 | 201 | |||
175↑ | 158↓ | 21↑ | 75↓ | 72↑ | 129↓ | |
↓ | ↓ | ↓ | ||||
Avg. sign. |
|
|
|
|||
>50 | 175↑ | 158↓ | 21↑ | 75↓ | 65↑ | 96↓ |
Fibroblasts incubated for 48 h with Claycomb medium, previously incubated for 24 h with cardiomyocytes were compared to fibroblasts incubated with fresh Claycomb medium.
Fibroblasts incubated for 48 h with supernatant from ultracentrifuged Claycomb medium, previously incubated for 24 h with cardiomyocytes were compared to fibroblasts incubated with fresh Claycomb medium.
Fibroblasts incubated for 48 h with pellet from ultracentrifuged Claycomb medium, previously incubated for 48 h with cardiomyocytes. Pellet was dissolved in DMEM and compared to fibroblasts incubated with fresh DMEM.
False Discovery Rate (FDR) was used for corrections for multiple testing. Significant up-regulation was defined as a foldchange >1.5 and significant down-regulation was defined as foldchange <0.67. A minimum signal intensity value of 50 was utilized. Abbreviations: Cm, cardiomyocytes; Fb, fibroblasts; sup. supernatant after ultracentrifugation; Avg. sign., average signal; FDR, False Discovery Rate; ↑, up-regulated; ↓, down-regulated; DMEM, Dulbecco's modified Eagle's medium.
We obtained microvesicles/exosomes released by murine cardiomyocytes. Microvesicles/exosomes were isolated from the culture medium by differential centrifugation and preparative ultracentrifugation. A corresponding maneuvre comprising the culture medium alone did not reveal any presence of microvesicles/exosomes. The isolated microvesicles/exosomes were subjected to transmission electron microscopy revealing small, rounded vesicles (40–300 nm) surrounded by a bilayered membrane. Some of them had a distinct electron dense appearance while others displayed an electron lucent interior (
A) Microvesicles/exosomes displaying an electron dense appearance, and B) electron lucent appearance. Bar represents 100 nm.
Microvesicles/exosomes prepared from Claycomb culture medium was incubated with antibodies conjugated with phycoerythrin (PE). A) Mouse anti-caveolin-3, was detected on approximately 30% of the microvesicles/exosomes. B) Mouse anti-flotillin-1, was detected on approximately 80% of the microvesicles/exosomes. C) Mouse anti-annexin-2, was not detected on the microvesicles/exosomes. D) Mouse anti-clathrin heavy chain, was not detected on the microvesicles/exosomes. The distribution of exosomes presenting caveolin-3 and flotillin-1 indicates that the sample contains more than one population of microvesicles/exosomes.
Since DNA unambiguously is present in prostasomes
A) Enhanced fluorescence at the 530±15 nm channel of membrane permeable acridine orange-stained microvesicles/exosomes (below) in comparison with unstained microvesicles/exosomes (above). B) Weak or no fluorescence at 670 nm/LP channel of membrane impermeable propidium iodide-stained microvesicles/exosomes (below) not differing from unstained microvesicles/exosomes (above).
We extracted total RNA from the microvesicles/exosomes. Genes/proteins (for which identified mRNAs are encoding) were used in the bioinformatic data base and a biological network was drawn (
Microvesicular/exosomal DNA and RNA were stained with AO. The AO-stained microvesicles/exosomes were thereafter dialysed for 24 h (with one change of dialysis buffer) and ultracentrifuged (130,000×g and 2 h at 4°C), to get rid of any trace amounts of irrelevant AO. Examination by confocal microscopy of fibroblasts incubated with AO-stained microvesicles/exosomes revealed AO-stained, intracellular DNA spots localized to and inside the nuclear membrane as well as spots of RNA (
Confocal microscopy picture of DNA-stained microvesicles/exosomes after dialysis, ultracentrifugation and resuspension in DMEM. After incubation with fibroblasts for 3 h at 37°C the DNA-staining localizes in fibroblasts to and inside the nuclear membrane. Additional light microscopy was used to add a layer in images to visualize cell borders. Arrows in A) and B) indicate acridine orange staining inside nuclei. B) also visualizes red wave length which detects acridine orange staining for RNA. Yellow staining shows colocalization of DNA and RNA.
When fibroblasts were transfected with microvesicles/exosomes a clearcut effect on gene expression was found. In more detail culture medium from cardiomyocytes induced 333 gene expression changes including 175 upregulations and 158 downregulations when compared to control medium. This culture medium gave rise to two compartments after ultracentrifugation: the microvesicle/exosome-deficient supernatant that induced only 96 changes (21 up, 75 down) and the microvesicle/exosome-enriched pellet resuspended in fresh DMEM that induced about 70% more changes (65 up and 96 down,
Parenchymal cells are believed to mould microenvironment components and affect various functions, mainly by pathways involving cell-to-cell contact and the release of soluble factors like autocrine products. However, an alternative novel mechanism that is now emerging involves the active release, as we show here, by cardiomyocytes of nucleic acid containing microvesicles/exosomes, with the capacity of complementary actions. Gupta and Knowlton
Apparently, DNA and RNA can be incorporated into microvesicles/exosomes and transferred to target cells here represented by fibroblasts. However, it remains elusive how the nucleic acids are sorted into microvesicles/exosomes. The microvesicle/exosome-linked nucleic acid may play a role in a variety of cell physiologic phenomena. Effective delivery of nucleic acids is crucial to their successful biological application. The microvesicles/exosomes appear to be naturally produced membraneous structures in the interstitial fluid with the ability to be recognized, adhered to and fused with other cell types, ultimately to deliver the cargo to target cells and their nuclei.
We confirm that microvesicles/exosomes derived from cardiomyocytes
Enzymes DNase I and II present in extracellular fluid (including blood plasma) degrade efficiently DNA and therefore only trace amounts of free (i.e. not shielded from an enzymatic attack) DNA can normally be detected in extracellular fluids. Accordingly, transfer of genetic information from one cell to another must be protected from DNase attack. We suggest that in case of cardiomyocyte intercellular communication this is accomplished by microvesicle/exosome-shielding of DNA.
The functional transfer of genetic information to target cells was in fact substantiated by clearcut down/up-regulation of gene products. This supports the idea that the content of microvesicles/exosomes (i.a. DNA, mRNA and proteins) was indeed internalised in fibroblasts rather than just adhered to them. This internalisation even into the nucleus was corroborated by the confocal microscopy findings.
In summary, we have demonstrated the presence of cardiomyocyte derived, nucleic acid-containing microvesicles/exosomes in media of cultured cardiomyocytes. We suggest that exosomes derived from cardiomyocytes could be denoted “cardiosomes”. Microvesicles and cardiosomes can be involved in a metabolic course of events in the microenvironment of the heart by facilitating an array of cellular processes through transfer of nucleic acids to target cells and their nuclei.
(TIF)
(TIF)
(TIF)
(DOC)
(DOC)
(DOC)
The authors acknowledge the able processing of electron micrographs by Berith Lundström and Professor Claycomb for supplying us with the HL-1 cell line.