| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,1
,1,¶,
,#,2,2,**,2
* Department of Medical and Surgical Sciences, University of Padova, Padova, Italy;
CORIT (Consorzio per la Ricerca sul Trapianto d’Organi), Padova, Italy;
Department of Biomedical Sciences, University of Padova, Padova, Italy;
CNR Institute of Neurosciences; Padova, Italy;
|| Institute of Pathological Anatomy, University of Padova, Padova, Italy;
¶ Clinica Chirurgica 3°, Padova General Hospital, Padova, Italy;
# Venetian Institute of Molecular Medicine, Padova, Italy;
** Direzione Sanitaria, Padova General Hospital, Padua, Italy
2Correspondence: S.S. and S.A.: Department of Biomedical Sciences, University of Padova, Via G. Colombo 3, Padova 35121, Italy. E-mail: stefano.schiaffino{at}unipd.it; ausoni{at}bio.unipd.it; E.C.: Clinical and Experimental Transplantation Immunology, Department of Medical and Surgical Sciences, University of Padova, Ospedale Giustinianeo, Via Giustiniani Padova, Italy. E-mail: emanuele.cozzi{at}unipd.it
ABSTRACT
Cardiomyocytes expressing host markers, such as the Y chromosome in sex-mismatched transplants, have been described in human allografts, suggesting that circulating cells can contribute to cardiac regeneration. It has not been established, however, whether host-derived cardiomyocytes result from transdifferentiation of stem cells or cell fusion. To address this issue, we used heterotopic heart xenografts and looked for markers of donor and recipient cells. Golden Syrian hamsters or transgenic mice expressing nuclear beta-galactosidase under the control of the cardiac troponin I promoter served as organ donors, while GFP+ transgenic rats were used as recipients. GFP+ cells, including abundant CD-45+ inflammatory cells and rare undifferentiated cells expressing early cardiac markers (GATA-4 or MEF2C), were found in xenografts harvested two weeks after surgery. In addition, rare GFP+ mature cardiomyocytes were found in 7 of 8 hamster xenografts and 6 of 6 mouse xenografts. The proportion of these cells was very low (0.0001% to 0.0344% in hamster xenografts) but similar to the one observed in control rat heart allografts. Without exception, all GFP+ cardiomyocytes also expressed donor markers, i.e., hamster membrane antigens or lacZ, so they must derive from cell fusion, not transdifferentiation.— Dedja, A., Zaglia T., Dall’Olmo, L., Chioato, T., Thiene, G., Fabris, L., Ancona, E., Schiaffino, S., Ausoni, S., Cozzi, E. Hybrid cardiomyocytes derived by cell fusion in heterotopic cardiac xenografts.
Key Words: stem cells • transgenic animals • cardiac transplantation
CARDIAC TRANSPLANTATION has been used as a model to explore the potential of circulating stem cells for recruitment to the damaged myocardium and their contribution to cardiac regeneration. Most studies have focused on human sex-mismatched heart transplants (female donor, male recipient), obtained at autopsy or as endocardial biopsies. Quaini et al. (1)
reported a high proportion (
18%) of cardiomyocytes containing the Y chromosome in the grafted hearts. This was interpreted as reflecting a marked cardiac chimerism resulting from the migration of circulating host cells into the transplanted heart and their rapid differentiation into cardiomyocytes. In a large number of subsequent studies, however, the Y-positive cardiomyocytes in similar sex-mismatched human heart transplants were reportedly in much lower proportion (ranging from 1.6% to 0) (2
3
4
5
6
7)
. The interpretation of these results is complicated by technical problems relating to Y chromosome detection by in situ hybridization (8
9
10)
, and the possibility of chimeric cells existing in the organ before transplantation, since they are also found in the hearts of normal women, due to blood transfusions or the transfer of fetal cells during pregnancy (11)
.
Using a heterotopic heart allograft model, in which normal rat hearts are transplanted into transgenic recipients expressing green fluorescent protein (GFP) in all tissues, we found a low proportion of cardiomyocytes expressing GFP (12)
. We were unable, however, to establish whether the presence of a host cell marker in the cardiomyocytes of the grafted heart is due to the transdifferentiation of circulating cells into cardiac muscle cells or the fusion of circulating cells with preexisting cardiomyocytes. To address this issue, we have now used two models of heterotopic cardiac xenotransplantation for which both donor and host cell markers are available. The results of these studies demonstrate that it is a process of fusion, not transdifferentiation, that is responsible for cardiac chimerism following heart transplantation.
MATERIALS AND METHODS
Animals
Wild-type (WT) adult Golden Syrian hamsters (n=8) and WT Sprague-Dawley (SD) rats (n=5) purchased from Harlan Italia (Italy), and transgenic mice (n=6) expressing beta-galactosidase under the control of the cardiac troponin I (cTnI) promoter (cTnI/lacZ, line 21) (13)
were used as heart donors. SD rats (n=19), expressing the enhanced GFP under the control of the cytomegalovirus (CMV) enhancer and the chicken beta-actin promoter (14)
, a gift from Dr. Okabe (University of Osaka, Japan), were used as cardiac recipients (15)
. All animals were kept in conventional facilities with free access to food and water. Adequate care for their health and well-being was provided in accordance with the Italian Animal Act (Law 116/92). These studies were conducted under the supervision of the Institutional Ethics Committee.
Surgical procedure, immunosuppression and postoperative follow-up
Splenectomy was performed first, followed by heterotopic abdominal cardiac xenotransplantation. The surgical procedure was undertaken according to the method already published by our group (16)
for the hamster-to-rat model and the technique used by Niimi (17)
, slightly modified at our center for the mouse-to-rat model. For the allografted control group, we used the technique described by Ono and Lindsey (18)
. Animals did not fast the night prior to surgery. Briefly, both donors and recipients were anesthetized with isoflurane (Forane, Abbott SpA, Campoverde di Aprilia, Italy) at 1–1.5% with oxygen, supplemented with 5 mg/kg of i.p. tramadol. After donor heparinization, hearts were harvested and flushed with Celsior (Imtix Sangstat, Lyon, France) cardioplegic preservation solution. The heart was immersed in cardioplegic solution and immediately transplanted under the operative microscope into the abdomen of the recipient. An end-to-side anastomosis was performed between the graft aorta and the abdominal aorta, and between the graft pulmonary artery and the inferior vena cava of the recipient. Fifteen mg/kg i.m. of cyclosporin A (Sandimmune, Novartis Italia, Rome, Italy) diluted in saline solution were administered to the recipient daily throughout the postoperative period, starting on day 0 (19)
. Analgesia with 5 mg/kg tramadol and adequate antibiotic coverage were used in the early postoperative period. Cardiac xenografts were monitored daily by direct palpation through the abdominal wall. Only animals with beating grafts at the time of cardiac harvest were included in the study.
Histopathological evaluation
Two weeks postoperatively (range 13–17 days), the grafts were removed and immediately fixed in 2% paraformaldehyde at room temperature for 2 h, equilibrated in sucrose gradient, and frozen in liquid nitrogen. Ten-micrometer sections from the midportion of the hearts were cut and stained with hematoxylin-eosin to assess tissue damage. The degree of tissue damage was scored according to a scale ranging from 0 to 2, in which grade 0 described no infiltrate, grade 1 a moderate perivascular and interstitial infiltrate with focal cell necrosis, and grade 2 a severe multifocal infiltrate accompanied by widespread tissue damage. All sections were analyzed using a Leica DMR optical microscope with a digital LEICA DC300 and IM1000 image software.
Immunofluorescence and confocal image analysis
GFP+ cardiomyocytes were detected and immunohistochemically characterized on 10-µm sections, using a panel of antibodies specific for alpha/beta myosin heavy chain (MyHC) (20)
, cTnI (21)
, laminin (rabbit antiserum, Sigma, Milan, Italy), rat CD-45 (30-F11 Pharmingen, Milan, Italy), Nkx2.5, GATA-4, MEF2C (N-19, H-112 and E-17, respectively, Santa Cruz Biotechnologies, Milan, Italy) and GFP (rabbit antiserum, Molecular Probes, Milan, Italy). All antibodies were diluted in 1% BSA, 0.2% Tween 20 in PBS and applied onto fresh cryosections at room temperature overnight. Reactivities were revealed by secondary TRITC-conjugated antisera (Dako, Milan, Italy and Jackson Laboratories, Soham, Cambridgeshire, UK). Some sections were counterstained with 4.6-diaminidino-2-phenylindole to visualize the nuclei. Confocal image analysis was performed using either the Bio-Rad MRC 1024ES Laser equipped with Nikon fluorescence and Bio-Rad software or the Ultraview Perkin Elmer confocal scanner equipped with the Nikon Eclipse TE200 microscope and the Imaging Suite 5.5 program.
Number of GFP+ cardiomyocytes
GFP+ cardiomyocytes were counted as explained elsewhere (12)
. Briefly, heart transplants were cryosectioned from the apex to the base. Two consecutive sections were analyzed every 100 µm. The first section was stained with an anti-laminin antibody (Ab) to count the GFP+ cardiomyocytes and the total number of cardiomyocytes in the section. The second section was stained with hematoxylin-eosin and used to measure the section area. The percentage of GFP+ cardiomyocytes in each transplant was obtained by dividing the total number of GFP+ cardiomyocytes in all sections by the total number of cardiomyocytes counted in the sections, obtained by multiplying cell density by section area.
Preparation of hamster-specific sera
Blood drawn from non-splenectomized, non-immunosuppressed rat recipients of a hamster cardiac xenograft was centrifuged at 3000 rpm for 10 min and stored in aliquots at –80°C. Each serum was diluted in PBS 1% BSA, centrifuged at 13000 rpm for 10 min and preabsorbed on fresh cryosections from rat heart for three consecutive steps of 20 min at 37°C. Preabsorbed sera were tested on heart cryosections incubated overnight at room temperature followed by 2 h of incubation at 37°C. Sections were washed in PBS and incubated for 40 min at 37°C with the anti-rat biotinylated secondary Ab (E0468, Dako, Milan, Italy), which was previously preincubated with rat non-immune serum for 5 min at 37°C. Reactivity with streptavidin-phycoerythrin (R0438, Dako, Milan, Italy) was revealed, according to the manufacturer’s instructions.
X-gal staining
To identify lacZ+ nuclei in cTnI/lacZ mice, 10-µm cryosections were incubated with 5'-bromo-4-chloro-3-indolyl-beta-D-galactopiranoside (X-gal, Inalco, Milan, Italy) at 37°C for 1 h, using standard procedures. This protocol enabled efficient beta-galactosidase detection without interfering with GFP visualization. Sections were analyzed by confocal microscopy using a Bio-Rad MRC 102ES Laser equipped with Nikon fluorescence and Bio-Rad software.
RESULTS
Two heterotopic cardiac xenotransplantation groups (hamster-to-rat and mouse-to-rat) were used to investigate the contribution of extracardiac circulating cells to cardiac regeneration. An additional allotransplantation model in the rat, with identical immunosuppressive treatment and splenectomy, was included as a control. In the heterotopic cardiac transplantation model the heart undergoes progressive atrophy due to unloading, but is well perfused and still beating at the time of graft retrieval. A variable degree of tissue damage is also observed as a consequence of transient ischemia during surgery and immune-mediated response, which is only partially reduced by immunosuppressive treatment. Recipients used in this study were GFP+ transgenic rats characterized by a strong, homogenous GFP expression in all tissues (12)
. Heart transplants were examined two weeks after surgery. In a previous study, we found a similar proportion of GFP+ cardiomyocytes in heart allografts retrieved 15 or 90 days after surgery (12)
.
Hamster-to-rat cardiac xenotransplantation
Macroscopic differences in heart size and strength of heartbeat were evident between grafts. Microscopic analysis confirmed a variable degree of tissue damage, inflammatory infiltration with granulation tissue and scar remodeling (Table 1
and Fig. 1
A, B). A large number of GFP+ cells engrafted the xenografts and were analyzed by confocal image analysis. Most of these cells stained positive for the CD-45 marker (not shown) and were, therefore, either inflammatory or stem cells of hematopoietic origin.
|
|
Antibodies to early cardiac markers, such as Nkx2.5, GATA-4 and MEF2C, were used to investigate the presence of GFP+ cardiac progenitors. Ab reactivity was first tested using fetal or neonatal rat heart cryosections, in which most of the cardiomyocytes stained positive (not shown). No expression of Nkx2.5 was ever seen in GFP+ cells in xenografts, whereas small undifferentiated GFP+ cells expressing GATA-4 (Fig. 1C-E
) or MEF2C (Fig. 1F-H
) were occasionally detected in the inflammatory infiltrate and/or close to normal cardiomyocytes in well-preserved myocardium. Positive cells, counted in 10 non-consecutive sections of three representative heart samples, were in the range of 0–3 (GATA-4) and 1–4 (MEF2C) per section. No small GFP+ cardiomyocytes coexpressing these markers with cardiac-specific contractile proteins were ever detected.
Seven of eight explanted hamster hearts revealed occasional GFP+ cells with the typical morphology and cross-striation of mature cardiomyocytes and staining for sarcomeric MyHC and cTnI (Fig. 2
A–D). These cells were mainly distributed within well-preserved myocardium, close to areas of tissue damage. They were mostly single cells, rarely double adjacent cells, never clustered cells. In hamster xenografts sectioned from the apex to the base, systematic GFP+ cardiomyocyte counts provided values ranging from 0.0001% to 0.0344%, comparable with those of the control allotransplantation group (Table 1)
. GFP expression by cardiomyocytes was confirmed by staining with an anti-GFP antibody (not shown). Ab reactivity was always confirmed by detection of native GFP, indicating that we never underestimated the number of positive cardiomyocytes due to any poor sensitivity of the detection method.
|
We used anti-hamster specific antisera to determine whether GFP+ cardiomyocytes in the transplants arose from the transdifferentiation of host-derived circulating cells or from the fusion of donor cardiomyocytes and host circulating cells. As in a previous report (22)
, we found that, following transplantation with a hamster heart, non-splenectomized and non-immunosuppressed rats develop a vigorous humoral anti-hamster immune response and produce Ab specific for hamster surface antigens. After initial immunohistochemical screening on hamster heart cryosections, we selected the three sera that provided the strongest reactivity. These sera were preadsorbed on rat tissue sections to remove cross-reactivity with rat antigens and then tested on normal hamster heart sections. As shown in Fig. 2E-H
, preabsorbed sera produced a bright stain at the surface membrane level in the hamster heart, and no staining in the control rat heart. In heterotopic heart transplants, all GFP+ cardiomyocytes observed in the grafts stained strongly at the membrane level with anti-hamster sera (Fig. 2I-J
). We never saw a single GFP+ cardiomyocyte failing to react to anti-hamster sera.
Mouse-to-rat cardiac xenotransplantation
In a second experimental model, adult cTnI/lacZ transgenic mice were used as organ donors. Histological analysis of mouse xenografts showed abundant fibrous scar tissue with a large number of small GFP+ cells corresponding to the inflammatory infiltrate (Fig. 3
A, B). As in the hamster xenografts, GATA-4 and MEF2C expression in small, undifferentiated cells was occasionally observed (Fig. 3C-H
).
|
To investigate the origin of mature GFP+ cardiomyocytes in the transplants, we took advantage of the fact that cTnI/lacZ21 transgenic mice express the lacZ marker exclusively in the heart (13)
. In this line, lacZ reporter is nuclear localized, and its expression is restricted to cardiac cells (Fig. 4
A, B). GFP+ cardiomyocytes were detected in each of the six grafts analyzed and expressed the typical repertoire of sarcomeric proteins (Fig. 4C, E
). Though not systematically quantitated in this model, the frequency of their occurrence was extremely low. Importantly, all the GFP+ cardiomyocytes identified in these xenografts coexpressed GFP and nuclear beta-galactosidase and were therefore hybrid cardiomyocytes containing both host and donor markers (Fig. 4C, D, F-H
). We found no small GFP+ cardiomyocytes coexpressing sarcomeric proteins and GATA-4 or MEF2C transcription factors (not shown), suggesting that cells identified by these markers do not progress toward cardiac maturation.
|
DISCUSSION
The main result of the present study is the demonstration that cardiac chimerism in grafted hearts is due to cell fusion and not cell transdifferentiation. This conclusion is supported by the analysis of two cardiac xenotransplantation models. In the first, in which hamster hearts were transplanted into GFP+ rats, we observed GFP+ cardiomyocytes in 7 of 8 grafts. The frequency of these cells was very low but was similar to the one previously observed in heart allografts (12)
and to the one found in the allogeneic control group in the present study. Our results indicate that, in the xenotransplantation model, circulating cells migrate and home spontaneously into the transplanted heart, and fuse with resident cardiomyocytes. That fusion, rather than transdifferentiation, is responsible for the formation of these cells is demonstrated by the finding that GFP+ cardiomyocytes were stained by a specific antihamster antiserum and thus coexpressed both donor and host cell markers. The coexistence of donor and host markers was also demonstrated in a genetic model, in which hearts from cTNI/lacZ transgenic mice were transplanted into GFP+ rats. The finding that all GFP+ cardiomyocytes, without exception, also stained for donor cell marker strongly supports the notion that cell fusion is the main process responsible for cardiac chimerism in experimental heart transplantation. The possibility of a transdifferentiation process also being initiated cannot be ruled out; indeed, a very limited number of small undifferentiated GFP+ cells expressing GATA-4 or MEF2C was found in both hamster-to-rat and mouse-to-rat grafts. These markers are not absolutely cardiac-specific, however, and, more importantly, we never detected any other cardiac-specific marker—such as cTnI or sarcomeric myosin—in these cells, or small GFP+ cardiomyocytes representing intermediate stages of maturation, which would suggest that GATA-4 or MEF2C positive cells never progress to become fully differentiated cardiomyocytes.
This paper provides the first demonstration of cell fusion in cardiac transplants. Cell fusion has been described in other settings, involving the injection of non-cardiac muscle cells into normal or injured myocardium or the bloodstream. Following the initial demonstration of cell fusion between bone marrow cells and cardiomyocytes in vitro (23)
and in vivo (24)
, cell fusion was subsequently demonstrated in other studies, including the transfusion of cardiac stem cells (25)
or bone marrow-derived hematopoietic cells (26
, 27)
into injured myocardium, the transplantation of skeletal muscle cells into normal hearts (28
29
30)
, the transfusion of human peripheral blood CD-34+ cells into mouse heart (31)
, and the coculture of human bone marrow stem cells with rat cardiomyocytes (32)
. Cardiomyocytes derived from cell fusion, and not transdifferentiation, were also recently observed in a xenogeneic context of human cord blood hematopoietic stem cells transplanted into neonatal immunodeficient mice (33)
.
The transdifferentiation vs. fusion issue is crucial. Given the low frequency of chimeric cardiomyocytes detected in most studies (both in heart transplants and after the injection of bone marrow cells), a precise knowledge of the mechanism involved in their formation is a prerequisite for procedures to be devised to enhance this process and thus exploit its therapeutic potential. That hybrid cell generation through cell fusion may have great potential and deserves serious consideration for cell therapy has been clearly demonstrated in a mouse model of hereditary tyrosinemia type I, in which mice with mutations in the fumaryl acetoacetate hydrolase (Fah) gene develop liver disease that can be corrected by the transplantation of WT bone marrow cells (34)
. It was subsequently shown that hepatocytes expressing Fah are not the product of a true transdifferentiation but arise from cell fusion between normal bone marrow cells and defective hepatocytes (35
, 36)
. The clinical improvement of the transplanted animals clearly demonstrates that cell fusion can produce functional cells, and this supports the notion that nuclear reprogramming occurs after cell fusion, in agreement with previous studies on heterokaryons (37)
, although an impaired donor cell gene expression has been observed in some cases.
At the moment, the most serious concern is the fact that spontaneous cell fusion is too rare to achieve significant clinical benefits unless a selective pressure is applied. In the Fah liver model, a strong selective pressure induced in vivo proliferation with formation of regenerating liver nodules by cell fusion events. This is unlikely to occur in the myocardium, so it would be important to devise procedures for enhancing the fusion process. The heterotopic heart transplantation model may offer some advantages in this respect. First, myocardial damage is consistently observed in the grafts and tissue damage is known to stimulate stem cell recruitment. Second, because the explanted heart can be maintained for several hours ex vivo in cardioplegic solution (12)
, it may be feasible to apply specific treatments before its reimplantation to identify factors able to enhance extracardiac cell recruitment and cell fusion or to promote the progression of transdifferentiation events, or to activate dormant endogenous cardiac stem cells. It is worth noting that insulin-like growth factor (IGF)-1 has recently been shown to enhance the fusion of bone marrow cells with adult skeletal muscle (38
, 39)
.
ACKNOWLEDGMENTS
This work was supported by the Consorzio per la Ricerca sul Trapianto d’Organi (CORIT), the University of Padua (Progetto di Ricerca di Ateneo 2005 CPDA054298 to S.A.), the Veneto Regional Authority, the Italian Ministry of Health and the European Union ("HeartRepair" LSHM-CT-2005–018630 to S.S.).
The authors thank Dr. Fabio Fante, Mr. Luigino Polito and Mr. Daniele Ramon for the breeding program and the veterinary care provided.
FOOTNOTES
1 These authors contributed equally to this work. 
Received for publication June 15, 2006. Accepted for publication August 7, 2006.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
