1. health and fitness
2. disease

Cardiovascular Research (2008) 77, 534–543
doi:10.1093/cvr/cvm071
Horizontal gene transfer from human endothelial cells to
rat cardiomyocytes after intracoronary transplantation
¨decke1, Alexander Assmann1, Andreas Wirrwar2,
Sandra Burghoff1, Zhaoping Ding1, Stefanie Go
2
3
¨ller2,
Doris Buchholz , Olga Sergeeva , Cordula Leurs4, Helmut Hanenberg4, Hans-Wilhelm Mu
5
1
¨rgen Schrader
Wilhelm Bloch , and Ju
1
Institute for Heart and Circulatory Physiology, Heinrich Heine University Duesseldorf, Universitaetsstr. 1, 40225 Duesseldorf,
Germany; 2Clinic for Nuclear Medicine, Heinrich Heine University Duesseldorf, Universitaetsstr. 1, 40225 Duesseldorf, Germany;
3
Department of Neurophysiology, Heinrich Heine University Duesseldorf, Universitaetsstr. 1, 40225 Duesseldorf, Germany;
4
Department of Pediatric Oncology Hematology and Immunology, Childrens Hospital, Heinrich Heine University Duesseldorf,
Universitaetsstr. 1, 40225 Duesseldorf, Germany; and 5Department of Molecular and Cellular Sport Medicine, German Sport
University, Cologne, Germany
Received 18 May 2007; revised 18 September 2007; accepted 24 October 2007; online publish-ahead-of-print 13 November 2007
Time for primary review: 21 days
KEYWORDS
Aims Recent studies suggested that human umbilical vein endothelial cells (HUVECs) transdifferentiate
into cardiomyocytes and smooth muscle cells in vitro. To test the functional relevance of this observation, we examined the transdifferentiation potential of HUVECs in vivo after intracoronary cell
application in Wistar rats.
Methods and results SPECT measurements (single photon emission computed tomography) revealed
that 18% of 111In-labelled HUVECs infused by intracoronary delivery stably transplanted to the rat
heart. For long-term tracking, HUVECs-expressing enhanced green fluorescent protein (EGFP) were
infused. Two days following transplantation, HUVECs were positive for caspase-3. Within 3 days, EGFP
was associated with individual cardiomyocytes. No labelling of endothelial and smooth muscle cells
was observed. The total number of EGFP-labelled cardiomyocytes accounted for 58% of all initially
trapped cells. These EGFP positive cells stained negatively for human mitochondrial proteins, but
were positive for rat monocarboxylate transporter-1 protein (MCT-1). Furthermore, EGFP-mRNA was
detected in these cells by single-cell RT–PCR (reverse transcription followed by polymerase chain reaction). After 21 days, EGFP positive cells were no longer observed. To investigate the underlying mechanism, we generated in vitro apoptotic bodies from EGFP-labelled HUVECs and found them to contain
the genetic information for EGFP. Co-incubation of apoptotic bodies with neonatal rat cardiomyocytes
caused cardiomyocytes to express EGFP.
Conclusion When transplanted into the rat heart by efficient intracoronary delivery, EGFP-expressing
HUVECs cause the exclusive but transient labelling of cardiomyocytes. Our in vivo findings suggest
that it is not cell fusion and/or transdifferentiation that occurs under these conditions but rather a horizontal gene transfer of the EGFP marker via apoptotic bodies from endothelial cells to cardiomyocytes.
Introduction
Ventricular remodelling after myocardial infarction and
subsequent development of heart failure cannot be prevented by the regenerative capacity of mature cardiomyocytes. Therefore, transplantation of beneficial cells is
thought to be a promising therapeutic approach. At
present, it is controversial whether a possible contribution
of these cells to functional improvements of the injured
* Corresponding
author. Tel: +49 211 8112671; fax: +49 211 8112672.
E-mail address: [email protected]
heart is based on transplant-related angiogenesis, paracrine
effects, or transdifferentiation into cardiomyocytes.1
Recent reports suggest that human umbilical vein endothelial cells (HUVECs) have the potential to transdifferentiate into heart-residing cells. For example, a significant
fraction of HUVECs were shown to differentiate into
smooth muscle like cells when the culture medium was
deprived of fibroblast growth factor, a process which can
be reversed by reapplicating the growth factor.2 Additionally, green fluorescent protein (GFP) expressing HUVECs
were reported to transdifferentiate into cardiac muscle
cells when cocultivated with freshly isolated neonatal
cardiomyocytes.3 The colocalization of endothelial proteins
Published on behalf of the European Society of Cardiology. All rights reserved. & The Author 2007.
For permissions please email: [email protected].
Downloaded from by guest on July 6, 2015
Cell therapy;
Transplantation;
Stem cells;
Catheter based stem cell
transplantation;
HUVECs;
Horizontal gene transfer;
Cardiac imaging;
Multipinhole-SPECT
Horizontal gene transfer from human endothelial cells to rat cardiomyocytes
dishes containing DMEM (Gibco), 20% foetal bovine serum (Biochrom
AG), penicillin G (100 U/mL), streptomycin (100 mg/mL) (Gibco).
After 2 h at 378C non-adherent cells were replated.
Apoptotic bodies were generated by incubation of HUVECs in
Basal Medium (PromoCell) without supplements for 24 h.9 Where
indicated,
0.5 mg/ml
ethidium
bromide
or
0.5 mg/ml
4,6-diamidino-2-phenylindole (DAPI, Sigma) was added. Medium
was then centrifuged for 10 min at 800 g, and the supernatant
again for 20 min at 16 000 g. Apoptotic bodies were resuspended
in complete medium.
For cocultivation assay conditions used were: 25 000 3-day old
neonatal rat cardiomyocytes were seeded onto sterile glas plates
(Ø22 mm, Plano, Wetzlar, Germany) and apoptotic bodies from a
100 mm flask of EGFP-expressing HUVECs were added. Cells were
incubated at 378C with fresh medium every 2 days. After 4 days,
cells were harvested and fixed for immunofluorescence staining.
Heart disintegration, single cell isolation, reverse
transcription, and PCR analysis
Hearts were perfused with 1 mg/ml Collagenase II (Biochrom AG)
and 5 mM CaCl2, minced and incubated at 378C for 15 min. Cells
were filtrated through a 100 mm Cell Strainer (BD Biosciences,
Heidelberg, Germany), centrifuged and resuspended in PBS. Green
fluorescent cells (excitation 485 nm, emission 530 nm) were
sucked into a sterile glass electrode (tip diameter 2–3 mm) filled
with 9 ml water. The content of the electrode was expelled into a
reaction tube containing 7 ml of a reverse transcriptase mixture
(Amersham Biosciences, Buckinghamshire, UK) and incubated at
378C for 1 h.
DNA from HUVECs and apoptotic bodies was isolated using the
DNeasy Tissue Kit (Qiagen GmbH, Hilden, Germany).
PCR was performed using primers for EGFP (50 -acgtaaacggccacaa
gttc-30 , 30 -cacatgaagcagcacgactt-50 ). Conditions for PCR amplification were: 30 s, 948C; 45 s, 578C; 75 s, 728C; 35 cycles.
Methods
The investigation conforms with the Declaration of Helsinki for use
of human tissue or subjects.
Cell culture and labelling
HUVECs were harvested by collagenase (Biochrom AG, Berlin,
Germany) and cultured to confluency in Basal Medium supplemented
with endothelial single quots (PromoCell, Heidelberg, Germany) on
gelatine-precoated culture flasks.
Second passage HUVECs (1 106) were incubated for 5 min with
7 MBq 111In (37 MBq/mL; Tyco Healthcare, Neustadt/Donau,
Germany) in serum free M199 medium (Gibco, Karlsruhe,
Germany) at room temperature. Labelling efficiency was measured
using a dose calibrator (MED Nuklear-Medizintechnik, Dresden,
Germany). Vitality of cells 90 min post-labelling was 66–84% (n = 4)
as tested by trypan blue exclusion.
For stable expression of EGFP, second passage HUVECs were transduced with lentivirus derived from the HIV1-vector pGJ37 expressing
EGFP under the control of the U3 promoter of spleen focus forming
virus, a mutant of the Friend mink cell focus-forming virus. Viral
particles were pseudotyped with the vesicular stomatitis virus glycoprotein G. HUVECs were washed (10 times) and passaged before
usage.
In vitro cocultivation assay
Neonatal rat cardiomyocytes were isolated from ventricular muscle
of newborn Wistar rats.8 Ventricles were minced in ice-cold PBS
buffer and incubated in 1% trypsin-EDTA (Sigma, Taufkirchen,
Germany) repeatedly at 378C for 15 min. Cell suspension was filtered through a 100 mm nylon mesh and seeded on 60 mm culture
Cyclosporine A experiments
HUVECs (1 104/cm2) were cultured on gelatine-precoated culture
flasks in Basal Medium supplemented with endothelial single quots
(PromoCell) and 6 mg/ml Cyclosporin (Sandimmunw, Novartis,
Nu
¨rnberg, Germany). Medium was changed daily. After 7 days cells
were fixed with 4% paraformaldehyde and 15% picric acid in PBS
and applied to immunohistochemistry.
Animal experiments
The investigation conforms with the Guide for the Care and Use of
Laboratory Animals published by the US National Institutes of
Health (NIH Publication No. 85-23, revised 1996). The intracoronary
transplantation of HUVECs was performed as previously described.6
Briefly, male Wistar rats (450–470 g) were intubated and anaesthetized by mechanical ventilation with isoflurane (1.5% v/v; Abbott,
Wiesbaden, Germany). After induction of transient cardiac arrest
[1 mM esmolol (Breviblocw, Gensia, San Diego, USA), 1 mM acetylcholin (Sigma) in 1 ml PBS] either 2.5 105 HUVEC (500 ml) or
MEM (500 ml, sham) were gently injected into the coronaries. All
animals received 30 mg/kg cyclosporine (Sandimmunw) the day
prior to transplantation and 15 mg/kg/d cyclosporine daily
thereafter.
For planar scintigraphic measurements during and after cell transplantation, a double headed gamma camera (PRISM 2000XP, Philips/
Picker, Hamburg, Germany; 10 frames for 15 s followed by 10 frames
for 5 min) equipped with a parallel hole collimator was used.
At indicated time points, animals were anaesthetized and echocardiography (Hewlett Packard Sonos 5500 equipped with 15 MHz
linear array probe) was performed.
Downloaded from by guest on July 6, 2015
with cardiomyocytic proteins within one cell was interpreted as evidence for transdifferentiation. Final proof by
using species-specific antibodies to discriminate between
transdifferentiation and cell fusion was not provided.3
Welikson et al.4 re-investigated these latter results, but
were not able to find transdifferentiated cardiomyocytes
originated from HUVECs. Only 1.2% of all cells were positive
for sarcomeric myosin and human nuclear antigen, but the
majority of these cells was binucleated suggesting cell
fusion between HUVECs and neonatal rat cardiomyocytes
rather than transdifferentiation. The direct injection of
HUVECs into the infarcted area of rat hearts led to an infiltration of macrophages and their subsequent phagocytosis.5
This was associated with improved left ventricular function
without signs of HUVEC transdifferentiation and might be
due to inflammatory responses.5
Given the divergent reports in the literature on endothelial cell plasticity, the present study explored the fate
of HUVECs under in vivo conditions in the beating rat
heart using a minimally invasive intracoronary delivery
system for cardiac cell transplantation.6 We found a large
fraction of the transplanted HUVECs having traversed the
coronary vascular wall and reached the interstitial space.
Thereafter, HUVECs became apoptotic and stained positive
for active caspase-3, whereas the infiltration of macrophages was not observed. Finally, solely rat cardiomyocytes
transiently expressed the HUVEC-introduced marker protein
enhanced GFP (EGFP). In our study, we did not find evidence
for cell fusion. Using species-specific antibodies, we ruled
out the possibility of transdifferentiation, but suggest that
horizontal gene transfer of EGFP through apoptotic bodies
explains our finding.
535
536
S. Burghoff et al.
Immunofluorescence and immunohistochemistry
Statistics
Data are expressed as mean + SEM. P , 0.05 was considered to be
statistically significant.
Total number of EGFP-expressing HUVECs (ntotal) was calculated
using the following formula:
ntotal ¼
nV
A ðd þ tÞ
where n is the number of counted cells per section, V the heart
volume, A the area analysed, d the diameter of HUVECs, and t
the thickness of the section.10,11
Results
Distribution of human umbilical vein endothelial
cells after intracoronary transplantation
For the intracoronary delivery of HUVECs into rats, a minimally invasive catheter-based technique was applied.6 The
position of the catheters during cell infusion is illustrated
in Figure 1A, and a representative recording of aortic
pressure during the intervention is shown in Figure 1B. In
order to visualize the distribution pattern of the cells
during and after intracoronary transplantation 111In-labelled
HUVECs (2.5 105 cells in 0.5 ml, 2.8 MBq) were transplanted. During cell delivery (within 30 s) and the following
50 min, planar views of the distribution of radioactivity
within the rat were obtained and permitted the calculation
of the number of cells which accumulated within the heart.
Fate of intracoronarely transplanted HUVEC
One hour after cell delivery, EGFP-expressing HUVECs were
found within the capillaries of the heart as evidenced by
endothelial staining for caveolin-1 and detected by epifluorescence (Figure 3A and B). This result was confirmed
when the staining was repeated using a rat specific
anti-PECAM-1 antibody (data not shown). To determine the
number of HUVECs within the heart, the number of green
fluorescent cells in 6 consecutive slices (each 7 mm) was
counted. The total number of all EGFP-expressing cells
within the heart was calculated to be 17% of all initially
transplanted HUVECs. To determine the origin of
EGFP-expressing cells in the rat heart, species-specific antibodies were used. One hour after transplantation of
EGFP-expressing HUVECs, green fluorescent cells stained
positively with anti-human mitochondria antibody (Figure
3C). No crossreaction with rat endothelial cells and rat cardiomyocytes was observed. The anti-rat MCT-1 antibody did
not stain EGFP positive HUVECs (Figure 3D). This antibody is
well suited to detect rat myocardium, since no immunostaining was observed after omission of the first antibody
(Figure 3E–G). One day after cell transplantation,
EGFP-expressing cells have crossed the endothelial barrier,
are still clearly positive for PECAM-1 and human mitochondrial proteins, and negative for rat MCT-1 (data not shown).
Three and seven days after transplantation of
EGFP-expressing HUVECs, all fluorescently labelled cells
showed typical cardiomyocytic morphology and appeared
to be well integrated in the myocardial tissue (Figure 4A).
No green fluorescent cells were detected in the endothelium
of rat coronary vessels. Careful examination of EGFP
Downloaded from by guest on July 6, 2015
Hearts were fixed with 4% paraformaldehyde for 12 h, rinsed in
0.1 M PBS for 3 10 min, stored for 24 h in PBS solution with 18%
sucrose and frozen at 2808C. Cryostat sections of 7 mm were cut
from cryoprotected cardiac tissue. Cells were fixed by applying
4% paraformaldehyde for 20 min and then rinsed in PBS for 24 h.
In case of immunohistochemical detection of EGFP samples
were incubated in 3% H2O2 for 30 min and thereafter in 0.5 M
NH4Cl, 0.25% Triton X-100 in 0.05 M TBS for 10 min. Blocking was
performed in 5% BSA in 0.05 M TBS for 1 h at room temperature.
Samples were marked with rabbit anti-human-mitochondria (pAB,
1:1000), chicken anti-rat-monocarboxylate transporter 1 (MCT-1)
(pAB, 1:1000), mouse anti-human-mitochondria (for doubleimmunohistochemistry, mAb, 1:1000) (Chemicon, Temecula, USA),
rat anti-mouse-CD31 (1:800, BD Pharmingen, San Jose, USA),
anti-caveolin-1 (pAB, 1:500, Transduction, Lexington, USA), mouse
anti-a-actinin (mAB, 1:400, Sigma), rabbit anti-GFP (pAB, 1:500,
Santa Cruz, CA, USA), or rabbit active-caspase-3 antibody (for
double-immunohistochemistry, BD Pharmingen, 1:500) in 0.8% BSA
in TBS overnight at 48C.
The secondary antibodies biotinylated goat anti-rabbit-IgG and
goat anti-mouse-IgG (1:400, Dako, Hamburg, Germany), biotinylated
sheep anti-rat-IgG (1:400, Amersham, LIFE SCIENCE, Little Chalfont,
UK), biotinylated goat anti-rabbit-IgG (1:400, Dako), and biotinylated goat anti-chicken-IgG (1:400, Promega, Madison, USA) were
used in 0.8% BSA for staining of samples. Extravidin Alexa 586 was
used for visualization of fluorescence of antibody binding. Where
indicated nuclei were counterstained with DAPI. Sections were analysed with confocal laser scanning microscopy (Zeiss, Oberkochen,
Germany). For immunohistochemistry detection of goat antirabbit-IgG occured with extravidin–horseradish-peroxidase-complex
(PRN1051, Amersham, 1:150, 60 min) and visualized by incubation
with DAB/NiSO4-solution (15 min). For subsequent detection of
goat anti-mouse-IgG streptavidin–alkaline-phosphate-complex
(D0396, Dako, 1:400, 60 min) was used and visualized by Fuchsin
substrate-chromagen system (Ko0624, Dako, 1 min).
As depicted in Figure 1C, radioactivity over the heart
reached a maximum of 36% of all cell associated radioactivity immediately after intracoronary cell infusion. Thereafter, radioactivity declined and reached a stable value
(halftime corrected) of 18% after 40 min. Figure 1D illustrates that cell distribution was site specific, since 111In
was found to be enriched over the heart. Similarly, threedimensional single photon emission computed tomography
(SPECT) data showed a specific and uniform accumulation
of labelled cells within the heart (data not shown).
Haemodynamic analysis revealed no significant changes in
ejection fraction and fractional shortening after HUVECs
transplantation (Figure 2).
In order to follow the cell fate for a longer period of time,
HUVECs were transduced with the lentivirus pGJ3-CSCGW
carrying the gene for EGFP.7 The endothelial character of
EGFP-expressing HUVECs was fully maintained up to nine
passages in culture as judged by the unchanged expression
of PECAM-1, VE-Cadherin, E-Selectin, and VCAM-1 after
induction with TNF-a (data not shown). Furthermore, EGFP
expression did not stimulate PGE2 release into the
medium, nor did it impair proliferation rate (data not
shown). For the in vivo experiments, viral stocks were
titrated to yield a transduction efficiency of maximal 30%
EGFP positive HUVECs, corresponding statistically to one
viral integration per transduced cell12 to ensure unchanged
endothelial phenotype. When these cells were transplanted,
EGFP positive cells were observed by epi-fluorescence
from endocardium to epicardium 7 days after cell delivery
(Figure 1E).
Horizontal gene transfer from human endothelial cells to rat cardiomyocytes
537
Downloaded from by guest on July 6, 2015
Figure 1 Intracoronary transplantation and distribution of human umbilical vein endothelial cells in the rat heart. (A) Schematic representation showing the
positions of extended arterial balloon (AB) in the aorta and venous balloon (VB) catheter in the right atrium of the rat heart. (B) Representative registration of
aortic pressure during cell infusion. 1, inflation of VB; 2, inflation of AB; 3, cardiac arrest by esmolol and ACh; 4, injection of human umbilical vein endothelial
cells; 5, deflation of AB and VB, injection of epinephrin and cardiopulmonary resuscitation; 6, normalization of aortic pressure. (C) Time course of radioactivity in
the heart during the delivery of 111In-labelled human umbilical vein endothelial cells. (D) Whole body distribution of human umbilical vein endothelial cells
45 min after transplantation. Bar = 1 cm. (E) Representative overview over the left ventricle from endocardium to epicardium showing an even distribution of
enhanced green fluorescent protein-expressing cells 7 days post-transplantation. Bar = 200 mm.
positive cells using laser scanning microscopy and epifluorescence detection revealed mononucleated cells in
investigated heart slices (7 mm), and mono- and binucleated
EGFP-expressing cells in dispersed cell preparation from
transplanted hearts (Figure 4A–C). Green fluorescent cells
stained positive for muscle protein a-actinin, thereby confirming a cardiomyocytic phenotype (Figure 4D). Using the
same calculation procedure as described above, the total
538
S. Burghoff et al.
To exclude the possibility that cyclosporine might have
altered the endothelial phenotype, HUVECs were cultured
for 7 days (n = 3) using culture medium that contained the
highest measured cyclosporine concentration in vivo [6 mg/
ml; 4 + 2 mg/ml (n = 3)]. After this period of time, HUVECs
were still positive for PECAM-1 and negative for a-actinin
and sm-actin (data not shown).
Apoptosis of human umbilical vein endothelial cells
and enhanced green fluorescent protein gene
transfer to rat cardiomyocytes
number of EGFP-expressing cardiomyocytes within the heart
after 7 days was estimated to be 58% of all HUVECs
trapped within the heart at 50 min after transplantation.
To define the species affiliation of EGFP-expressing cells, 3
and 7 days after transplantation immunofluorescence
revealed that green fluorescing cells were also positive for
rat MCT-1 protein (Figure 4E), but did not stain for human
mitochondrial proteins (Figure 4F), whereas the antibody
is well suited to detect human cardiac cells (Figure 4H). In
order to demonstrate EGFP transcripts in EGFP-expressing
rat cardiomyocytes, single cell PCR on such cells was performed. To this end, cells were separated from a heart
that had received EGFP-expressing HUVECs 6 days ago, isolated single EGFP positive cardiomyocytes, and performed
reverse transcription followed by PCR. As shown in the
representative Figure 4G, EGFP-expressing rat cardiomyocytes display a positive signal demonstrating the existance
of EGFP-mRNA within rat cardiomyocytes, whereas control
cardiomyocytes did not. Interestingly, we found dot like
staining for human mitochondrial proteins ubiquitiously distributed and aligned to the cellular membrane of cardiomyocytes (Figure 4I), a phenomenon not observed in sham
operated rats (inset Figure 4I).
Expression of EGFP in rat cardiomyocytes was transient.
Three weeks after transplantation of labelled HUVECs, no
EGFP-expressing cells were detected within the heart
using epi-fluorescence. Also, there was no longer a positive
dot like staining of the anti-human mitochondria antibody
after this period of time (data not shown).
Discussion
This study investigated the in vivo transdifferentiation
potential of HUVECs. To this end, we transplanted 111In or
EGFP-labelled HUVECs via the coronary arteries into rat
hearts and found that 18% of the infused HUVECs were
retained evenly distributed within the heart immediately
Downloaded from by guest on July 6, 2015
Figure 2 Ejection fraction and fractional shortening of rat hearts after
transplantation of enhanced green fluorescent protein-expressing human
umbilical vein endothelial cells over 3 weeks. Rats were measured under isofluoran anaesthesia before operation (n = 8), after operation (n = 8) and at
time points of sacrifice (n = 2 for days 1, 3, 7 and n = 1 for day 21). During
operation either enhanced green fluorescent protein-expressing human umbilical vein endothelial cells (filled square) or medium (filled circle) were intracoronarely infused resulting in no significant functional changes.
In order to study the mechanism of disappearance of HUVECs
in vivo, hearts were excised 2 days after cell transplantation
and analysed for apoptosis by immunohistochemistry. As
shown in Figure 5A and B, only hearts that had received
HUVECs contained caspase-3 activated cells. Figure 5C and
D shows two representative apoptotic cells stained for
active-caspase-3 (black) and human mitochondrial proteins
(fuchsin-red). As can be seen cells are double positive and
show an apoptotic morphology.
Since apoptotic bodies have been reported to be capable
of transferring genetic information to target cells13,14 we
next prepared apoptotic bodies from EGFP-expressing
HUVECs with DAPI and ethidium bromide-labelled DNA and
evaluated in a first step their composition (Figure 6A–D).
Isolated apoptotic bodies contained nucleic acids, but did
not contain visible amounts of EGFP. PCR analysis confirmed
that apoptotic bodies from EGFP-expressing HUVECs, but not
from non-transduced HUVECs, carry the gene for EGFP
(Figure 6E).
To investigate whether apoptotic bodies can be taken up
by cardiomyocytes studies with rat neonatal cardiomyocytes
were performed. Coincubation of rat cardiomyocytes with
apoptotic bodies derived from EGFP-expressing HUVECs for
4 days clearly revealed EGFP-expressing cells as shown
using immunohistochemistry (Figure 6F and G). These cells
are positive for rat MCT-1 (Figure 6H–K, epi-fluorescence)
and myosin-heavy-chain but negative for human mitochondrial proteins (data not shown). Even after incubation of
apoptotic bodies with DNAse, in order to exclude the possibility of EGFP plasmid transfer, rat MCT-1 positive cells were
detected (data not shown). Gene transfer was a rather rare
event: we found 8–10 EGFP positive cells per culture plate.
It should be noted that despite optimized culture conditions
only cardiomyocytes that incorporated the label lost their
typical morphology and viability, whereas EGFP negative
cardiomyocytes in the same culture plate appeared structurally unaltered. To exclude viral shuttle to cardiomyocytes,
supernatant from HUVECs was added to HEK293T and
HT1080 cells, which are easily transduced by viruses, and
incubated for 5 days. FACS analysis did not reveal any
green fluorescing cells (data not shown) indicating the
absence of active virus in the supernatant of our HUVEC
cultures.
Horizontal gene transfer from human endothelial cells to rat cardiomyocytes
539
Downloaded from by guest on July 6, 2015
Figure 3 Localization of enhanced green fluorescent protein-expressing human umbilical vein endothelial cells immediately after intracoronary application as
detected by epi-fluorescence. (A) Enhanced green fluorescent protein-expressing human umbilical vein endothelial cells (green) are shown in rat myocardium.
(B) Visualization of capillary endothelium by staining against caveolin-1 (red, visualized with Alexa 586) demonstrates intravascular localization of the enhanced
green fluorescent protein positive cells. (C ) Immunostaining with anti-human mitochondria antibody (red, visualized with Alexa 586) reveals the human origin of
enhanced green fluorescent protein positive cells by overlay of enhanced green fluorescent protein and Alexa 586 fluorescence (yellow). (D) Rat specific
anti-MCT-1 antibody (red, visualized with Alexa 586) detected only rat cardiomyocytes, whereas enhanced green fluorescent protein positive cells were negative.
(E–G) The secondary antibody alone did not stain rat myocardium. Shown is an enhanced green fluorescent protein-expressing human umbilical vein endothelial
cell in rat heart tissue where enhanced green fluorescent protein is green (E), secondary antibody is red (F) and (G) shows both images merged. Bars = 5 mm.
after transplantation. After 2 days, HUVECs underwent
apoptosis and after 3 days EGFP was exclusively associated
with rat cardiomyocytes. Using species-specific antibodies
and confocal microscopy, we excluded HUVECs transdifferentiation and cell fusion with cardiomyocytes as the underlying cause. We rather provide evidence that apoptosis and
apoptotic bodies generated from transplanted HUVECs
were responsible for directly transmitting nucleic acids
coding for EGFP to rat cardiomyocytes.
Numerous reports suggested that endothelial cells
are able to transdifferentiate into different cell types such
as smooth muscle cells and myocytes. Although transdifferentiation of human endothelial progenitor cells into
cardiomyocytes appears to be an extremely rare event,15
these cells give rise to skeletal muscle cells when transplanted into mice.16 In contrast, in coculture experiments
with neonatal rat cardiomyocytes, no convincing evidence
of transdifferentiation into cardiomyocytes was obtained.17
Mature bovine aortic endothelial cells are also able to transdifferentiate into smooth muscle cells in vitro.18 Similarly,
HUVECs can form smooth muscle-like cells after cultivation
without fibroblast growth factor.2 In the present study, we
found no evidence for transdifferentiation of HUVECs
under in vivo conditions, despite an efficient transfer of
EGFP from HUVECs to cardiomyocytes. Species-specific antibodies did not detect human protein within the target cells,
540
S. Burghoff et al.
Downloaded from by guest on July 6, 2015
Figure 4 Fate of enhanced green fluorescent protein-expressing human umbilical vein endothelial cells 7 days after intracoronary application. Detection of
enhanced green fluorescent protein positive (green) and of Alexa 586 fluorescing cells (red) in rat myocardium 7d after human umbilical vein endothelial cell
transplantation (A–F, I ) and in human myocardium (G). (A–C ) Merged image of enhanced green fluorescent protein positive cells either on heart sections (A)
or dispersed cells (B, C) and 4,6-diamidino-2-phenylindole staining of the nuclei (blue) show the typical shape of cardiomyocytes and their nuclei (arrows).
(D) a-Actinin staining revealed a cross-striation like pattern in enhanced green fluorescent protein positive cells. (E, F ) enhanced green fluorescent protein positive cells are stained by rat specific anti-MCT-1 antibody (E), whereas there was no immunostaining against human mitochondrial proteins in enhanced green
fluorescent protein positive cells (F ). (G) Single cell PCR using primer against enhanced green fluorescent protein on enhanced green fluorescent
protein-expressing cardiomyocyte, enhanced green fluorescent protein negative cardiomyocyte, enhanced green fluorescent protein-expressing human umbilical
vein endothelial cell and water as control. The amplified fragment contains 187 bp. (H ) Clear mitochondrial staining of a specimen from human heart, using
anti-human mitochondria protein antibody. (I ) Dot like immunostaining (arrows) in the interstitium of the rat heart 7 days after human umbilical vein endothelial
cell transplantation when stained with anti-human mitochondria antibody (I ). In the heart of sham operated rats, no specific staining with anti-human mitochondria antibody appears (inset Figure 1). Bar (A) = 50 mm, (B)–(I ) = 20 mm.
Horizontal gene transfer from human endothelial cells to rat cardiomyocytes
541
rather all EGFP positive cardiomyocytes were of rat phenotype. It should be noted that HUVECs are much smaller
than cardiomyocytes and transdifferentiation would have
required considerable cell growth within the short period
of 3 days in addition to being well integrated within the
texture of the heart. Another possible explanation for
EGFP-expressing rat cardiomyocytes is fusion of HUVECs
with cardiomyocytes. In accordance with data in the literature,19,20 we found rat ventricular cardiomyocytes often to
be binucleated and to some extent mononucleated. Similarly, we found mono- and binucleated cells to be green
fluorescent both in heart slices and dispersed cardiomyocytes. In all cases however, we did not detect human
mitochondrial proteins within green fluorescent cardiomyocytes, definitely ruling out the possibility of cell fusion.
We always noted that EGFP-expression in cardiomyocytes
exceeded that of HUVECs despite the fact that the single
volume of cardiomyocytes is by far higher than that of
HUVECs suggesting amplification of the EGFP message
within the target cell. Indeed, we were able to detect EGFPspecific mRNA within green fluorescent rat cardiomyocytes.
This clearly shows the transfer of nucleic acids coding for
EGFP from HUVECs to rat cardiomyocytes in vivo.
The transfer of proteins from one cell type to another
might also be achieved by nanotubular connections
between cells.21–23 In our study, we did not find microscopic
evidence for nanotubes between HUVECs and cardiomyocytes neither in vitro nor in vivo. If nanotubes were at all
involved in the transfer of label in our experiments, this
would have required the transport of nucleic acids since
we detected mRNA and observed signal amplification in
the target cell. Yet, nanotubular transport of nucleic acids
has not been reported so far.
Blomer et al.24 observed a viral shuttle from lentivirally
transduced cardiomyocytes to recipient fibroblasts. Viral
shuttle took place when cocultivation was started immediately after viral transduction and infectiosity was lost
when the virus was incubated at 378C for several hours
before transduction and after passaging the cells. In contrast, our cells were passaged once and kept in the incubator for several days in addition to extensive washing steps
and still showed a transduction rate which is unlikely to
result from direct viral shuttle. In addition, we did not
find evidence for infectiosity when supernatants from
EGFP-expressing HUVECs were cultured with HEK293T or
HT1080 cells. This makes it unlikely that viral particles
Downloaded from by guest on July 6, 2015
Figure 5 Immunostaining of activated-caspase-3 and human mitochondrial proteins in rat hearts 2 days after intracoronary application of medium or human
umbilical vein endothelial cells. (A) Control hearts perfused with cell free medium did not stain against activated-caspase-3. (B) Hearts which received
human umbilical vein endothelial cells show activated-caspase-3 positive cells, indicating apoptosis. (arrows) (C, D) Two representative double-immunochemical
images of hearts after human umbilical vein endothelial cell infusion with antibodies against human mitochondrial proteins (violet-red) and active-caspase-3
(black) providing evidence that transplanted human umbilical vein endothelial cells undergo apoptosis in vivo. The cell in (C ) shows a progressed state of apoptosis. Bars (A) and (B) = 100 mm, (C ) and (D) = 50 mm.
542
S. Burghoff et al.
Downloaded from by guest on July 6, 2015
Figure 6 Genetic information for enhanced green fluorescent protein-expression is transferred to rat cardiomyocytes via apoptotic bodies. DNA from enhanced
green fluorescent protein-expressing human umbilical vein endothelial cells and non-transduced human umbilical vein endothelial cells was labelled with ethidium bromide and apoptotic bodies were isolated. (A–D) show apoptotic bodies stained with 4,6-diamidino-2-phenylindole (A) and ethidium bromide (B) in
brightfield (C ) and all images merged (D). Bar = 5 mm. (E) Gel electrophoresis of PCR products from DNA isolated from enhanced green fluorescent
protein-expressing human umbilical vein endothelial cells, non-transduced human umbilical vein endothelial cells, and their respective apoptotic bodies
(apt.B.) to detect enhanced green fluorescent protein. (F–G) Chromogenic detection of an enhanced green fluorescent protein-expressing neonatal rat cardiomyocyte after addition of apoptotic bodies derived from enhanced green fluorescent protein-expressing human umbilical vein endothelial cells 4 days ago (F ) and
control of non-expressing cells (G). Bar = 50 mm. (H–K ) The same experiment as in (F ) showing neonatal rat cardiomyocytes-expressing enhanced green
fluorescent protein (green, H ), staining positive for rat MCT-1 (red, I ) and both images merged (K ). Bar = 10 mm.
adhered to HUVEC membranes and became incorporated
into apoptotic bodies. Secondly, we did not achieve stable
EGFP-expression in recipient cells, which argues against
virus shuttle as well.
The partial transformation of apoptotic cells into
phosphatidylserine-containing apoptotic bodies has been
reported earlier.25 In cell culture, apoptotic bodies were
shown to increase the number and differentiation state of
endothelial progenitor cells.9 In a study by Holmgren
et al.,14 the cellular uptake of DNA from apoptotic bodies
resulted in the transfer of the foreign genomic DNA to the
nucleus of the phagocytosing cell and the expression of
the marker gene at the protein and mRNA level. Even
whole chromosomes or fragments thereof have been
reported to be transferred by this pathway.13 Our data
strongly suggest that the in vivo transfer of the EGFP label
to cardiomyocytes was preceded by apoptosis of transplanted HUVECs. Not only contained HUVECs activatedcaspase-3 as an apoptotic signal 2 days after transplantation
which were absent 1 day later, we also found that all
EGFP-expressing cells were rat cardiomyocytes at that
time point. Thus, our in vivo and in vitro data are in
support of the view that apoptotic bodies derived from
EGFP-expressing HUVECs transferred the genetic information for EGFP to cardiomyocytes. Since the expression
of EGFP does not persist, we assume that the transferred
DNA does not integrate into the chromosomes of the recipient cell and is subsequently removed. Should additionally
mRNA be taken up, degradation of the transferred material
can be expected. Also note that not all genes are equally
transferred, since we did not find human mitochondrial proteins in rat cardiomyocytes. Whether this phenomenon is
restricted to virus-inserted genes is presently unclear.
EGFP has been widely used in the literature for stable
cell-labelling in order to follow cell migration and/or
phenotypic changes. Certainly, without additional control
experiments, the sole association of the EGFP label with a
cell of different phenotype provides no evidence for the
Horizontal gene transfer from human endothelial cells to rat cardiomyocytes
mechanism involved. Although in the past, transdifferentation and cell fusion were considered to be the main alternatives in studies with endothelial cells, the present study
shows that horizontal gene transfer via apoptotic bodies is
an important third alternative to be considered.
Conflict of interest: none declared.
Funding
Forschungskommission of the Heinrich Heine University
Duesseldorf (9772228 to S.B.), DFG (SFB612 to J.S.).
References
11. Weibel ER. Stereological Methods Vol. 2: Theoretical Foundations.
London: Academic Press; 2007.
12. Kustikova OS, Wahlers A, Kuhlcke K, Stahle B, Zander AR, Baum C et al.
Dose finding with retroviral vectors: correlation of retroviral vector copy
numbers in single cells with gene transfer efficiency in a cell population.
Blood 2003;102:3934–3937.
13. Bergsmedh A, Szeles A, Henriksson M, Bratt A, Folkman MJ, Spetz AL
et al. Horizontal transfer of oncogenes by uptake of apoptotic bodies.
Proc Natl Acad Sci USA 2001;98:6407–6411.
14. Holmgren L, Szeles A, Rajnavolgyi E, Folkman J, Klein G, Ernberg I et al.
Horizontal transfer of DNA by the uptake of apoptotic bodies. Blood 1999;
93:3956–3963.
15. Badorff C, Brandes RP, Popp R, Rupp S, Urbich C, Aicher A et al. Transdifferentiation of blood-derived human adult endothelial progenitor cells
into functionally active cardiomyocytes. Circulation 2003;107:
1024–1032.
16. Pesce M, Orlandi A, Iachininoto MG, Straino S, Torella AR, Rizzuti V
et al. Myoendothelial differentiation of human umbilical cord
blood-derived stem cells in ischemic limb tissues. Circ Res 2003;93:
e51–e62.
17. Gruh I, Beilner J, Blomer U, Schmiedl A, Schmidt-Richter I, Kruse ML
et al. No evidence of transdifferentiation of human endothelial progenitor cells into cardiomyocytes after coculture with neonatal rat cardiomyocytes. Circulation 2006;113:1326–1334.
18. Frid MG, Kale VA, Stenmark KR. Mature vascular endothelium can give
rise to smooth muscle cells via endothelial-mesenchymal transdifferentiation: in vitro analysis. Circ Res 2002;90:1189–1196.
19. Li F, Wang X, Gerdes AM. Formation of binucleated cardiac myocytes in
rat heart: II. Cytoskeletal organisation. J Mol Cell Cardiol 1997;29:
1553–1565.
20. Li F, Wang X, Bunger PC, Gerdes AM. Formation of binucleated cardiac
myocytes in rat heart: I. Role of actin-myosin contractile ring. J Mol
Cell Cardiol 1997;29:1541–1551.
21. Koyanagi M, Brandes RP, Haendeler J, Zeiher AM, Dimmeler S. Cell-to-cell
connection of endothelial progenitor cells with cardiac myocytes by
nanotubes: a novel mechanism for cell fate changes? Circ Res 2005;96:
1039–1041.
22. Onfelt B, Nedvetzki S, Yanagi K, Davis DM. Cutting edge: Membrane nanotubes connect immune cells. J Immunol 2004;173:1511–1513.
23. Rustom A, Saffrich R, Markovic I, Walther P, Gerdes HH. Nanotubular
highways for intercellular organelle transport. Science 2004;303:
1007–1010.
24. Blomer U, Gruh I, Witschel H, Haverich A, Martin U. Shuttle of lentiviral
vectors via transplanted cells in vivo. Gene Ther 2005;12:67–74.
25. Henson PM, Bratton DL, Fadok VA. Apoptotic cell removal. Curr Biol 2001;
11:R795–R805.
Downloaded from by guest on July 6, 2015
1. Srivastava D, Ivey KN. Potential of stem-cell-based therapies for heart
disease. Nature 2006;441:1097–1099.
2. Ishisaki A, Hayashi H, Li AJ, Imamura T. Human umbilical vein
endothelium-derived cells retain potential to differentiate into smooth
muscle-like cells. J Biol Chem 2003;278:1303–1309.
3. Condorelli G, Borello U, De Angelis L, Latronico M, Sirabella D, Coletta M
et al. Cardiomyocytes induce endothelial cells to trans-differentiate into
cardiac muscle: implications for myocardium regeneration. Proc Natl
Acad Sci USA 2001;98:10733–10738.
4. Welikson RE, Kaestner S, Reinecke H, Hauschka SD. Human umbilical vein
endothelial cells fuse with cardiomyocytes but do not activate cardiac
gene expression. J Mol Cell Cardiol 2006;40:520–528.
5. Merx MW, Zernecke A, Liehn EA, Schuh A, Skobel E, Butzbach B et al.
Transplantation of human umbilical vein endothelial cells improves left
ventricular function in a rat model of myocardial infarction. Basic Res
Cardiol 2005;100:208–216.
6. Ding Z, Fach C, Sasse A, Godecke A, Schrader J. A minimally invasive
approach for efficient gene delivery to rodent hearts. Gene Ther 2004;
11:260–265.
7. Leurs C, Jansen M, Pollok KE, Heinkelein M, Schmidt M, Wissler M et al.
Comparison of three retroviral vector systems for transduction of nonobese diabetic/severe combined immunodeficiency mice repopulating
human CD34+ cord blood cells. Hum Gene Ther 2003;14:509–519.
8. Pinson A. Cell Culture Techniques in Heart and Vessel Research. Berlin:
Springer-Verlag; 1990.
9. Hristov M, Erl W, Linder S, Weber PC. Apoptotic bodies from endothelial
cells enhance the number and initiate the differentiation of human endothelial progenitor cells in vitro. Blood 2004;104:2761–2766.
10. Weibel ER. Stereological Methods Vol. 1: Practical Methods for Biological
Morphology. London: Academic Press; 2007.
543