Cardiac Allograft Vasculopathy after Cardiac Transplantation and Hormone Therapy: Positive Effects?

Abstract

Background. 

There is a great deal of controversy surrounding the issue of hormone replacement therapy after transplantation. The question whether or not this therapy has effects in cardiac allograft vasculopathy (CAV), the Achilles heel of cardiac transplantation or other unique aspects of allograft function is still unknown.

Methods. 

We investigated the long-term effect of 17β-estradiol as well as phytoestrogen Coumestrol, a synthetically produced phytoestrogen, on the development of CAV and the degree of fibrosis in an ovariectomized female heterotopic chronic allograft model (LEW-F344).

Results. 

We found that, 150 days after transplantation, no significant effect of estrogen application on intimal thickening of coronary arteries was observed. 17β-estradiol and phytoestrogen Coumestrol did significantly reduce the perivascular immune reaction. However, the immune effect had no consequence on the intensity of CAV. Although neither 17β-estradiol nor phytoestrogen Coumestrol revealed a positive effect on CAV, the group of animals treated with 17β-estradiol showed the highest decline in heart function and the most distinct fibrosis.

Conclusions. 

17β-estradiol does not affect CAV positively, but worsens cardiac allograft function and leads to increased fibrosis. This is the first study showing a negative effect of 17-β-estradiol after heart transplantation in the long term.

Although in recent years improved immunosuppression could reduce events of acute cardiac allograft rejection, chronic rejection still remains a severe clinical problem for the long-term graft survival. Overall 5-year survival for cardiac transplantation is 65% with a mortality rate of 4% per year from year 1 to year 14 (¹). Chronic heart rejection is characterized by a progressive, obliterative myointimal hyperplasia, also known as cardiac allograft vasculopathy (CAV). The concentric intimal thickening is mainly composed of T-lymphocytes, macrophages, and smooth muscle cells, together with major histocompatibility complex (MHC) class II presenting endothelium (^{2, 3}). The mechanisms behind transplant arteriosclerosis have not yet been elucidated in detail and although experimental and clinical studies have suggested that CAV is a T-cell-dependent event (^{4, 5}), it presumably also involves nonimmunologic factors (⁶). To date, there is no effective treatment to prevent arteriosclerosis in heart transplants.

Years ago, various studies investigated the effects of estradiol on vascular injury. In this context it was shown that it inhibits proliferation of smooth muscle cells (^{7, 8}), myointimal thickening and DNA synthesis (⁹). In models of rabbit heart and rat aortic transplant arteriosclerosis treatment with estradiol was shown to prevent myointimal hyperplasia (^{10, 11}) and abolish the upregulation of MHC-class II expression in the coronary arteries of rabbit cardiac allografts (¹²). Clinically, the physiological production of estrogen is thought to protect young women against arteriosclerosis. However, postmenopausal hormone therapy (HT) has never been approved for the prevention of coronary heart disease (CHD) due to its various side effects and recently, an increased risk for CHD with use of HT was observed (¹³). Additionally, in heart transplantation, female sex negatively affects short-term graft survival (¹⁴). In the present study, we wanted to explore whether exogenous estradiol in a model using female ovariectomized rats would affect the progression of CAV. We were also interested in the effect of phytoestrogens since they lack a modulating impact on the reproductive system.

In the present study we applied the LEW to F344 rat heterotopic heart transplantation model to investigate the long-term effect (150d) of 17β-estradiol and the phytoestrogen Coumestrol on the development of CAV, the degree of fibrosis and the reaction of the immune system in ovariectomized female rats.

MATERIALS AND METHODS

Inbred female LEW (RT^avl) and F344 rats (RT1^lvl) were obtained from Charles River Laboratories (Sulzfeld, Germany), ovariectomized F344 rats (RT1^lvl) from IFFA CREDO (l‘Arbresle, France). Before the surgical procedure, animals were acclimatized for ≥7 days with controlled light/dark cycles and ad libitum access to food and water. Principles of laboratory animal care were followed according to the current version of the “German Law on the Protection of Animals.”

Heterotopic Heart Transplantation

Syngenic (F344 → F344), as well as allogenic (LEW → F344) heterotopic heart transplantations were performed according to the method of Ono and Lindsey (¹⁵) with modifications (¹⁶). In brief: donor rats (12 weeks old) were anesthetized with Enfluran via inhalation. Cardiac arrest was done by injecting cold (4°C) Bretschneider-cardioplegic solution (Custodiol, Dr. F. Köhler Chemie, Alsbach-Hähnlein, Germany) first retrograde into the vena cava inferior, followed by an antegrade perfusion of the coronary arteries via the aorta. An atrial septum defect was set via the right atrium and the leaflets of the tricuspid valve were resected to generate insufficiency. After ligation of vessels, the heart was carefully excised. The recipients (F344, ovariectomized, 12 weeks old) were also anesthetized with Enfluran via inhalation. The allograft was transplanted by anastomosing the aorta end to side to the abdominal aorta and the right atrium to the vena cava inferior of the recipient.

Experimental Groups

F344 rats with syngrafts (syn; n=7) and those with allografts (allo; n=11) received a 14-day course of cyclosporine A (Novartis, Nuernberg, Germany) 2 mg·kg⁻¹·day⁻¹ i.m., only. The estradiol treated group (E; n=7) received additionally 2 μg·kg⁻¹ of 17β-estradiol (Sigma-Aldrich, Taufkirchen, Germany; dissolved in NaCl) s.c. daily. The phytoestrogen treated group (PE; n=8) received 2 mg·kg⁻¹ of Coumestrol (Sigma, Schnelldorf, Germany; dissolved in 100% ethanol and administered in peanut oil) i.m. twice a week, and the oil+ethanol group (n=8) peanut oil and ethanol i. m. twice a week. Medication started 14 days prior to surgical procedure and lasted until harvest of the transplant at day 150 (Table 1).

TABLE 1: Experimental Design

Graft function was assessed through daily abdominal palpation and scored on a scale of 0–4 (0=absence of beating; 4=optimal beating). Then 150 days after surgical procedure all animals were weighed and transplants, native hearts and uteri were explanted. Atrial tissue was removed and the ventricles were divided in two parts. One part was fixed in 3.5% formalin (stabilized with methanol) overnight and embedded in paraffin. The other part was placed in tissue freezing medium (Leica Instruments GmbH, Nussloch, Germany) and immediately shock frozen in liquid nitrogen.

Histology

For each heart, paraffin sections (2 μm) were stained with hematoxylin and eosin to investigate morphology. Elastica von Gieson (EVG) staining was applied to identify fibrosis. The degree of fibrosis was scored in myocardial tissue and perivascular area by light microscopy at ×125 or ×500 magnification on a scale of 0 to 3, as follows: 0=negative (i.e. same as native hearts), 1=weakly positive, 2=moderately positive, 3=extremely positive.

Morphometric Analysis

The degree of intimal thickening was determined by computer-assisted morphometry on sections stained with EVG. The outer border of the intima is thereby visualized by the red stained internal elastic lamina. The sections were scanned with an Axioplan microscope through a Neofluar 25× objective (NA=0.6; Zeiss, Oberkochen, Germany) coupled to a digital camera (DXM 12000; Nikon, Japan). The image size consisted of 640×512 pixels, each pixel in 24 bit (RGB). The spatial resolution of the scanning system was 1.6 pixel per μm. The beam of the illumination source (100W, halogen lamp) was filtered by a blue filter for contrast enhancement in the measured image. The image analysis was performed on a workstation XP1000 (Compaq, USA) by algorithms programmed in FORTRAN. Vessels not obliquely cut or at a branching point, as well as vessels <15 μm or >90 μm in diameter were not used for analysis. One point (one pixel) within the lumen of each vessel was marked interactively in the image seen on the screen of the workstation. The area surrounding this point up to the internal elastic lamina was detected automatically by computer algorithms using color and geometrical information (¹⁷). The degree of transplant vasculopathy is expressed as a percentage of the area of intimal thickening to the total vessel area.

Immunohistochemistry

Deep-frozen (−80°C) tissue was cryosectioned (4 μm) with a cryostat (Frigocut N, Leica), mounted on slides coated with 3-aminopropyltriethoxysilane and air dried followed by fixation in acetone for 10min at room temperature. The slides were rinsed 3× in 138.5 mM Tris-NaCl, pH 7.4 and incubated with monoclonal antibodies against CD4 (OX38; 1:800), CD8 (OX8; 1:800), macrophages/monocytes (ED1; 1:1000) or MHC class II (anti-IA; 1:1000) at room temperature for one hour in a humid chamber. Monoclonal antibodies were obtained from BIOZOL Diagnostica (Eching, Germany). Diluted normal rat serum and 1.5% hydrogen peroxide served to block unspecific staining. After repeated rinsing in Tris-NaCl the sections were incubated with HRP-conjugated rabbit anti-mouse IgG (1:50; Dako Diagnostica GmbH, Hamburg, Germany) at room temperature for 30min in a humid chamber, rinsed again in Tris-NaCl and incubated with HRP-conjugated goat anti-rabbit IgG (1:50, Medac, Wedel, Germany) at room temperature for 30min in a humid chamber. Slides were rinsed 3× in Tris-NaCl, then put in PBS, pH 7.4 and specific antibody binding was detected with DAB (Sigma). Following immunostaining slides were counterstained with hematoxylin using standard techniques and coverslipped with Kaiser's glycerol gelatin (Merck KGaA, Darmstadt, Germany). Specific antigen expression was determined in myocardial tissue and in perivascular area in a blinded manner by light microscopy at x125 or x500 magnification and was graded on a scale from 0–3, as follows: 0=negative (i.e. same as native hearts), 1=weakly positive, 2=moderately positive, 3=extremely positive.

Statistical Analysis

An unpaired Student's t test was used for statistical analysis (SPSS, SPSS Inc., Chicago, IL) of CAV, body weight, and uterus weight data. Results are expressed as mean ± standard error of the mean (SEM), P<0.05 was considered statistically significant.

A Mann Whitney U test (STATISTICA 7.1: StatSoft Inc. Tulsa, OK) was used for the scored data. Results are expressed as median and interquartile range in parenthesis (IQR), P<0.05 (two tailed) was considered statistically significant.

RESULTS

Of the 41 heterotopic heart transplants performed, no graft was lost due to acute rejection, all grafts had palpable contractions at the time of explantation and were visibly pulsatile at harvest. Immediately after transplantation all groups had good graft function, scored on a scale of 0–4 (median (IQR)): syngenic 3.5 (2.5–4.0), allogenic 4.0 (4.0–4.0), estradiol 4.0 (3.5–4.0), phytoestrogen 3.75 (3.25–4.0) and oil+ethanol 4.0 (4.0–4.0). At day 150 (day of harvest) graft function was worse in all groups (syngenic 2.5 (2.5–3.0), allogenic 3.0 (3.0–3.5), estradiol 2.5 (2.0–3.0) (E vs. allo: P<0.05), phytoestrogen 3.25 (2.5–3.75) and oil+ethanol 3.75 (2.5–4.0); however, allografts treated with 17β-estradiol showed the highest loss in beating score (Table 2).

TABLE 2: Graft function immediately after transplantation and at the day of harvest in heart syngrafts and allografts

Independent of treatment all animals gained weight following transplantation. Body weight of rats receiving estradiol increased significantly less compared to those with only CsA (37.3±5.4 g (E) vs. 81.5±4.4 g (allo), P<0.001; Fig. 1), whereas uterine weight was significantly higher (384.4±19.7 mg (E) vs. 57.0±5.5 mg (allo), P<0.001; Fig. 1).

FIGURE 1.: Increase in body weight (hatched bars) and uterine weight (closed bars) after 150 days. F344 rats with syngrafts (syn) and with allografts (allo), receiving CsA only. F344 rats with allografts, treated with 17β-estradiol (E), with the phytoestrogen Coumestrol (PE) or with peanut oil and ethanol (oil+ethanol). Results are presented as means ± SEM *P<0.05 treated vs. CsA-only group.

Histology

Fibrosis could be detected in myocardial tissue and perivascular area in all groups, with lowest values in syngrafts (Table 3). Myocardial tissue was always more affected than the perivascular area. In contrast to allografts receiving Coumestrol or those treated with peanut oil and ethanol, 17β-estradiol treated grafts revealed significantly (P<0.05) more fibrosis when compared to those with CsA only (Table 3).

TABLE 3: Fibrosis in heart syngrafts and allografts

Morphometric Analysis

In allografts, 150 days posttransplantation 28.8±1.3% of the vessel area was obstructed, compared to 15.3±0.8% in syngrafts (P<0.001; Figure 2). 17β-estradiol did not significantly affect intimal proliferation (26.0±1.9%). Application of the phytoestrogen Coumestrol significantly reduced intimal thickening (21.5±1.2%, P<0.001; Figure 2). However, this could also be achieved by the application of peanut oil and ethanol only (21.9±1.2%, P<0.001).

FIGURE 2.: Intimal thickening in arteria of syn- or allografts with CsA-only (allo), or treated with 17β-estradiol (E), the phytoestrogen Coumestrol (PE), or peanut oil + ethanol. Intimal thickening is expressed as a percentage of myointimal area to total vessel area (=area within the internal elastic lamina). Results are presented as means ± SEM; #P<0.05 allo- vs. syngrafts; *P<0.05 treated vs. CsA-only group.

Immunohistochemistry

T lymphocyte invasion could be observed in both, syn- and allografts in myocardial tissue as well as in perivascular area (Table 4). In myocardial tissue, positive immunostaining for CD4⁺ and CD8⁺ T lymphocytes was less in the treated (E, PE) than in the CsA-only group, however, without reaching statistical significance. In contrast, T lymphocyte invasion into perivascular area was significantly reduced by application of 17β-estradiol (CD4⁺: 1.0 (1.0–1.0) vs. 2.0 (2.0–2.0), P<0.01; CD8⁺: 1.0 (1.0–1.0) vs. 2.0 (1.0–2.0), P<0.01; Table 4) or Coumestrol (CD4⁺: 1.0 (1.0–1.5) vs. 2.0 (2.0–2.0), P<0.01; CD8⁺: 1.0 (1.0–1.0) vs. 2.0 (1.0–2.0), P<0.05; Table 4).

TABLE 4: CD4⁺ and CD8⁺ T lymphocytes in myocardial tissue or perivascular area of heart syngrafts and allografts

Immunostaining identified macrophages/monocytes in myocardial tissue and in perivascular area in all groups (Table 5). 17β-estradiol treated allografts had significantly less macrophages/monocytes than those with CsA-only in both myocardial tissue and perivascular area, respectively (1.0 (1.0–2.0) vs. 2.0 (2.0–2.0), P<0.01 and 1.0 (1.0–1.0) vs. 2.0 (1.0–2.0), P<0.05, respectively; Table 5). Application of Coumestrol affected macrophages/monocytes comparably.

TABLE 5: Macrophages/monocytes in myocardial tissue or perivascular area of heart syngrafts and allografts

Cells expressing MHC class II antigen could be detected in myocardial tissue and in perivascular area in all groups (Table 6). In myocardial tissue neither 17β-estradiol- nor coumestrol-treatment altered antigen expression significantly. In perivascular area, antigen expression was significantly reduced by 17β-estradiol (2.0 (2.0–2.0) vs. 3.0 (3.0–3.0), P<0.01; Table 4) and coumestrol (2.0 (1.5–2.0) vs. 3.0 (3.0–3.0), P<0.01; Table 6).

TABLE 6: MHC class II antigen expression in myocardial tissue or perivascular area of heart syngrafts and allografts

DISCUSSION

The hypothesis that estrogen protects the vasculature from injuries and arteriosclerosis was supported by the lower incidence of coronary artery disease in premenopausal compared to postmenopausal women or men. In contrast, immune responses are enhanced in women compared with men (¹⁸) and following organ transplantation rejection episodes are more frequent in female than in male recipients (¹⁹). Estrogen might hence be responsible for the increased risk of rejection, but could also be inhibitory for the development of transplant arteriosclerosis. The value and effects of estradiol replacement therapy in female transplant recipients is thus unknown and could be either beneficial or detrimental. However, following the recent publication of negative effects of hormone therapy in postmenopausal women this hormone replacement therapy is no longer widely recommended (¹³).

In the present study, a well-established experimental model was used to help elucidate the long-term effects (150 days) of estrogen therapy on the development of CAV in heart transplantation in the absence of endogenous hormone production by using female donors and ovariectomized female recipients. All other previously performed experimental studies in this context used male rabbits for heterotopic heart transplantation or male rats for transplantation of aortic segments or cardiac allografts. The LEW-F344 combination is a model requiring CsA therapy only for a limited period and showing an ongoing low-grade rejection process with development of graft arteriosclerosis.

In this model, we did not observe a significant effect of estrogen application on intimal thickening of coronary arteries 150 days after transplantation. Though we did not determine estrogen levels in the blood, the dramatically higher uterus weight of rats in the group receiving 17β-estradiol should be sufficient to proof therapy. Dose of 17-β-estradiol was due to previous studies of a group at the University Clinic of Wuerzburg dealing with heart insufficiency (Pelzer T. et al., unpublished data). In these studies, an estradiol dosage was evaluated that equals physiological estradiol synthesis in rats. Additionally, in these studies the effect of various serum levels of estradiol on body weight and uterus weight was investigated.

Foegh and coworkers demonstrated, in various experimental studies in the 1980s and 1990s, positive effects of estrogen therapy in different animal models (^8–12). In this context, the application of 100 μg/kg/d of 17β-estradiol in a heterotopic cardiac allograft model did significantly inhibit myointimal hyperplasia six weeks after transplantation (¹⁰). However, this effect was shown with male rabbits receiving a diet high in cholesterol and supraphysiologic levels of estrogen and the development of myointimal hyperplasia was investigated already 6 weeks after transplantation (²⁰).

The desired positive effects of estrogen will always be hampered by its negative ones on the reproductive system. However, since the discovery of the estrogen receptor (ER) subtype β, which was shown to be predominant in rat cardiac allografts and to be responsible for the vasculoprotective effects of estrogen (²¹), phytoestrogens have gained in significance. Phytoestrogens have a higher binding affinity to ERβ than to ERα (²²) and could thus provide vasculoprotection without any uterotrophic side effects.

Most investigations were performed with the phytoestrogen genistein, a phytoestrogen from soy protein (²³). Soy protein favorably affects all serum lipoprotein concentrations. It decreases serum cholesterol, LDL-cholesterol, and triglyceride levels significantly while increasing serum HDL-cholesterol concentrations slightly (²⁴). Soy isoflavones have antioxidant properties that protect LDL from oxidation (²⁵). They may decrease platelet aggregation and thus decrease the tendency for thrombosis of blood vessels, the major cause of heart attacks and strokes (²⁶). Additionally soy, isoflavones have favorable effects on blood vessel function (²⁷).

In the present study, Coumestrol, a synthetically produced phytoestrogen, was applied. In the group receiving Coumestrol, intimal thickening of coronary arteries was significantly less compared to that without Coumestrol. However, for application, Coumestrol has to be dissolved in oil and ethanol and therapy with oil and ethanol alone showed the same effect. Hence, it can be assumed that Coumestrol, applied in dosages of 2 mg/kg twice a week, does not affect CAV. Interestingly, soy protein fed to monkeys with the phytoestrogens intact showed a positive effect on lipid metabolism and the reduction of atherosclerosis. However, also soy protein with the phytoestrogens extracted revealed positive effects compared to the control group (²⁸). There are hints that cardioprotective agents like antioxidants are more effective when given in their natural environment than as extracts or synthetic products. It can thus not be excluded that natural phytoestrogens, e.g. in soy protein, would positively affect CAV.

Although estrogen did not reveal a negative effect on CAV, we observed worse heart function at termination of the experiment in the 17β-estradiol treated group. Concomitantly this group showed the most distinct fibrosis, interstitial as well as perivascular. The capacity of estrogen to stimulate neonatal rat cardiac fibroblast has already been shown in vitro (²⁹).

Previous studies postulated an inhibitory effect of estrogen on the immunosuppressive property of cyclosporine. Zou et al. observed a reduced cardiac allograft survival in a mouse model after ovariectomy and estradiol treatment (³⁰). The authors concluded that an inhibitory effect of estrogen on the immunosuppressive property of cyclosporine A is responsible for the enhanced acute rejection. Survival was prolonged by the estrogen-antagonist tamoxifen. However, the inhibitory effect of estrogen on the immunosuppressive property of cyclosporine might not be that effective in our weak allogen rat strain combination. Nevertheless, because we did not look for acute rejection, the effect of early immunosuppression inhibition and the loss in graft function can not be excluded.

17β-estrogen as well as Coumestrol did significantly reduce the perivascular immune reaction in the long term, however without any consequence on the intensity of CAV. In contrast, Koskinen et al. and Heemann et al. demonstrated that a reduction of CD4+ (and /or CD8+) positively affects CAV (^{31, 32}).

In summary, neither 17β-estradiol nor the phytoestrogen Coumestrol revealed a positive effect on CAV in our heterotopic transplantation model. Surprisingly, the group receiving 17β-estradiol showed the highest decline in heart function and the most distinct fibrosis. This would be interesting to be investigated in detail. The question remains if the ovariectomy with loss of physiological hormone production can be compensated by the application of synthetically produced estrogen and phytoestrogen.

Whether the positive effects on CAV seen in the group receiving peanut oil and ethanol is due to components in peanut oil or ethanol or both has to be investigated in further studies.

Based on our results, we do not have any hints for a reduction of CAV with hormone therapy. However, the hormone dose applied in ovariectomized rats does not correspond to human hormone therapy. Nevertheless, together with the results in postmenopausal women, the use of hormones for cardioprotection seems questionable.

ACKNOWLEDGMENTS

The authors thank Prof. A.M. Waaga (Department of Surgery, University of Wuerzburg, Wuerzburg, Germany) and Prof. A. Chandracker (Transplantation Research Center, Brigham and Women′s Hospital and Children′s Hospital Boston, Harvard Medical School, Boston, MA for their kind support; R. Wahn and E. Grella (Department of Cardiac and Thoracic Surgery, University of Wuerzburg, Wuerzburg, Germany) for invaluable technical assistance.