Morphologic Study of Homograft Valves before and after Cryopreservation and after Short-Term Implantation in Patients

Morphologic Study of Homograft Valves before and after Cryopreservation and after Short-Term Implantation in Patients Y. A. H. Goffin, MD, PhD,* R. Henriques de Gouveia, MD,* † T. Szombathelyi, MD,*‡ M. J. M. Toussaint, MSc, PhD,§ and E. Gruys, DVM, PhD§ *European Homograft Bank, Brussels, Belgium; †the Department of Pathology/Morphology, Hospital de Santa Cruz, Lisbon; ‡the Department of Forensic Medicine, Semmelweis Medical School, Budapest; §the Department of Pathology, Faculty of Veterinary Medicine, University of Utrecht, Utrecht.

11 Cryopreserved heart valve homografts have been implanted in patients for the past 15 years, but controversies still exist on the survival of donor cells, matrix maintenance, and possible rejection by the host. Therefore a full morphologic study (histology, immunohistochemistry, transmission electron microscopy, and cuprolinic blue-TEM for glycosaminoglycans [GAG]) of short-term implanted uninfected grafts was done using unimplanted valves as the reference. Unimplanted tissues consisted of 5 fresh and 11 cryopreserved valves. Eight implants were recovered at reoperation (4) or autopsy (4), 4 from the right and 4 from the left ventricular outflow tract. The implantation time was 2 hours to 30 days. For unimplanted valves we found a partial preservation of the endothelium, the presence of dendritic Langerhans cells (Lc) and macrophages, and no significant damage to fibroblasts, collagen framework, and GAG pattern, except when the tissues had been ischemic for a long time. Explanted cusps exhibited (i) early disappearance of endothelium and Lc; (ii) nonspecific low-grade inflammatory cell infiltration, mostly of monocytoid type; (iii) viable degree of devitalization of fibroblasts with persistence of viable cells in some areas in most cusps; and (iv) fair preservation of collagen framework and GAGs. It is likely that, in view of the good graft preservation at implantation, humoral rejection is responsible for the earlier destruction of the endothelium and dendritic cells and the delayed devitalization of the fibroblasts and that preservation of the collagen framework and other intercellular matrix components (glycosaminoglycans) should guarantee longterm graft function. © 1997 by Elsevier Science Inc. Cardiovasc Pathol 1997;6:35–42

Since 1975 the clinical use of a new generation of heart valve homografts, namely, cryopreserved valves, has become increasingly popular among cardiac surgeons. So far, few long-term follow-up results on this new generation of valves are available (1–4), the longest term series consisting of a 10-year study by O’Brien et al. published in 1987 (3) and whose findings were confirmed by a recent 15year follow-up report (4). In the first study O’Brien et al. (3) compared the behavior of “fresh nonviable” aortic valves

Manuscript received January 29, 1996; revised July 22, 1996; accepted July 25, 1996. Address for correspondence: Professor Y. A. H. Goffin, European Homograft Bank, c/o Military Hospital, Rue Bruynstraat, 1120 Brussels, Belgium; Phone: 32-2-264-40-66, fax: 32-2-268-51-71. Cardiovascular Pathology Vol. 6, No. 1, January/February 1997:35–42  1997 by Elsevier Science Inc. 655 Avenue of the Americas, New York, NY 10010

with their so-called “viable” cryopreserved counterparts and reported that the actual freedom from reoperation at 10 years was significantly higher in the latter group. This was confirmed by their 15-year results reporting a freedom from reoperation of 79%. In their 10-year study the authors also reported the pathologic findings in three cryopreserved aortic valves, explanted at 2, 10, and 20 months, respectively. They observed histologic evidence of apparently viable fibroblasts in certain cusps or cuspal areas. However, no mention was made of the endothelial covering of the valves and the presence or absence of Langerhans cells (Lc), which are both highly immunogenic. With regard to the immunogenicity of homograft valves, clinical observations either suggest (5) or give evidence (6,7) of a variable degree of early and transient humoral rejection after implantation of either antibiotic preserved “homovital”

1054-8807/97/$17.00 PII S1054-8807(96)00072-5

36

GOFFIN ET AL. CRYOPRESERVED HOMOGRAFT VALVES

Cardiovasc Pathol Vol. 6, No. 1 January/February 1997:35–42

or cryopreserved valves (5,6) without repercussion on valve function. According to some authors the intensity of the cytoimmunologic response seems to depend on the preservation method of the homografts (7) and the donor–recipient blood group system matching (6). More recently, histological observations made by Schoen et al. could not confirm the finding of viable cells in explanted cryopreserved aortic valves even after short-term implantation, in contrast with the persistence of viable cells in the valves of cardiac allografts submitted to an immunosuppressive regimen (8). It has become evident from the explant studies that further research must be done to determine whether one should transplant homografts with viable cells. The present work has been undertaken to (i) identify the cellular populations present in the valves before and after implantation, namely, endothelial, connective tissue, dendritic Langerhans, or possible inflammatory cells; (ii) evaluate the state of preservation of the collagenous framework and ground substance (glycosaminoglycans [GAGs]) of the cusps; and (iii) detect the possible effects of the patients’ early immune reaction toward the implants. The findings made in explanted grafts are referred to by cell types and state of preservation of these cells and their matrix in two categories of unimplanted homografts: noncryopreserved and cryopreserved and thawed.

Material and Methods Material Twenty-four unimplanted tissues were studied: (i) 5 cryopreserved aortic valve homografts were submitted to immunohistochemical cell identification; (ii) 4 noncryopreserved (2 aortic and 2 pulmonary) and 11 cryopreserved valves (7 aortic and 4 pulmonary) were processed for transmission electron microscopy (TEM) to study the state of preservation of both the cells and the matrix. A total of 8 implanted homograft valves removed from 8 patients were studied: 5 aortic and 3 pulmonary. All homografts were prepared from the heart of beating heart donors. The age of the graft recipients ranged from 2 days to 64 years.

In 6 cases the graft had been implanted in an anatomic position, and in a heterotopic position in the 2 others (1 pulmonary in the left ventricular outflow tract [LVOT] and 1 aortic in the right VOT). Explants with infective endocarditis and partial explants devoid of cusps were not included in the study. The implantation duration varied from 2 hours to 30 days postoperation. 4 valves were removed at autopsy and four explanted at surgery. The reason for surgical explantation was technical failure in three cases and tissue failure in one (a case of early postoperative cusp rupture considered to be the result of incorrect thawing, see Table 1).

Routine Histology All samples were fixed either in 4% formaldehyde or in a buffered solution of 4% formaldehyde, 1% glutaraldehyde. Paraffin sections were stained with hematoxylin-eosin with or without alcoholic saffron. In certain cases sections were subjected to an elastin stain or to alcian blue at pH 2.5 for proteoglycans.

Immunohistochemical Characterization of Cell Populations An avidin-biotin-peroxidase (ABC) technique was carried out, according to an original protocol of the laboratory of anatomic pathology of Cliniques Universitaires St Luc, Brussels (Pr. J. Rahier): Thin sections were deparaffinized and endogenous peroxidases blocked. In formalin-overfixed tissues, the microwave heating process (antigen retrieval system) was then used to recover antigenicity. Inhibition of aspecific reactions was performed by incubation with normal goat serum 1⁄10 1 1% of bovine serum albumin for 30 minutes at 378C. A panel of primary antibodies (ab) included three murine monoclonal ab: antivimentin (dilution 1;50; Enzo USA), antismooth muscle actin (dilution 1;1000; Dako), and anti-CD68 (dilution 1;50, Dako; and three polyclonal rabbit ab: anti-factor VIII (dilution 1;1000; Dako, anti-S100 protein (dilution 1;12,000; Dako) and anti-CD3 (dilution 1;2000; Dako). The secondary ab used were anti-mouse (rat monoclonal 1;500; Boehringer-Mannheim) and anti-rabbit (polyclonal 1;500;

Table 1. Short-Term Implants (Listed According to Implantation Duration) Case No. 1 2 3 4 5 6 7 8

Implantation

Valve Type

Donor Category

Sex/Age

Site

Duration

Graft Removal

Valve Function

P A P (mono cusp) P A A A A

bh bh (CO intoxication) bh bh bh bh bh bh

F/50 M/22 M/22 M/51/2 M/64 M/60 M/2 days M/35

LVOT LVOT RVOT RVOT LVOT LVOT RVOT LVOT

2 hours 8 days 10 days 10 days 10 days 20 days 21 days 30 days

Surgery Autopsy Surgery Autopsy Autopsy Surgery Autopsy Surgery

Tissue failure: cusp rupture* Normal function Technical failure: stenosis distal anastomosis Normal function Normal function Technical Failure: paravalvular leak Normal Function Technical failure: regurgitation on geometric mismatch

Abbreviations: A 5 aortic valve, bh 5 beating heart, LVOT 5 left ventricular outflow tract, P 5 pulmonary valve, RVOT 5 right ventricular outflow tract. *Thawing accident.

Cardiovasc Pathol Vol. 6, No. 1 January/February 1997:35–42

37

GOFFIN ET AL. CRYOPRESERVED HOMOGRAFT VALVES

Table 2. Unimplanted Cryopreserved Valves: Immunohistochemistry of the Cell Population of the Cusps Antibody Reactions Donor Case Valve Sex/ No. Type Age 1 2 3 4 5

A A A A A

M/13 F/14 M/51 M/53 M/60

Clinical Diagnosis Dilated CMP* Restrictive CMP Coronary artery disease Atherosclerosis; dilated CMP Atherosclerosis; brain hemorrhage

Actin Vimentin (Smooth Muscle) prot S100 (Lcs) (Mesodermal Cells) ann. fib. sp. v. ann. fib. sp. v. 111 11 11 11 11

2 6 6 1 1

2 2 2 2 2

2 2 2 6 2

2 2 2 6 2

11 11 111 111 11 2 11 6 11 2 11 1 11 2 11 2 11 2 11 2

Factor VIII (Endothelium)

CD 68 (Macrophages)

6 1 1 11 1

11 2 6 1 1

Abbreviations: A 5 aortic valve, ann 5 annulus, CMP 5 cardiomyopathy, fib 5 fibrosa, sp 5 spongiosa, v 5 ventricularis, 2 5 negative, 6 5 scarce positive cells, 1 5 a few, 11 5 fair number , 111 5 numerous. *Also recent past history (4 months) of parvovirus myocarditis.

Boehringer-Mannheim). Streptavidin-POD-conjugate (1;1000; Boehringer-Mannheim) was then applied. Final staining was done with diaminobenzidine (DAB). Sections were counterstained with Mayer’s hematoxylin. Positive controls consisted of a lymph node, a cutaneous hemangioma, and the root tissues of the valves (internal controls). The intensity of staining for the antigens was semiquantitatively recorded as 2 (negative), 6 (scarce positive cells), 1 (a few), 11 (fair number), and 111 (numerous).

Transmission Electron Microscopy Samples were fixed in a Dulbecco buffered solution of 4% formaldehyde and 1% glutaraldehyde. Then the tissue samples were rinsed in 0.1 M cacodylate buffer at pH 7.2 and postfixed with 2% OsO4. Staining was done with 2% uranyl acetate. Durcupan-embedded thin sections were poststained with lead citrate. For GAG staining a second set of tissue samples was kept overnight in a 2.5% glutaraldehyde solution containing 0.025% cuprolinic blue and 0.2 M MgCl2 buffered with 0.025 M sodium acetate. After rinsing in buffered glutaraldehyde and staining with 1% sodium tungstate, the specimens were dehydrated in acetone. Embedding was in durcupan. To control the types of GAGs, a series of valves was incubated with chondroitinase ABC, chondroitinase AC, and nitrous acid according to the method of Van Kuppevelt et al. (9) before incubation.

Results Unimplanted Valves Reminder of the histologic features of pulmonary and aortic valves. The cusps consist essentially of three components: endothelial cells, which are often irregularly present on the valvular surfaces, collagen fibers, and elastic tissue. The architecture consists of three layers that are distinctly present from the annulus to the linula where they become

undistinctive: a dense fibrous layer on the arterial surface, the fibrosa, an elastin-rich layer on the ventricular surface in the continuation of the infundibular endocardium, the ventricularis, and an intermediate layer of loose connective tissue, the spongiosa. The annulus of each cusp has a double-layered structure with a fibrous annulus close to the sinus bottom and an underlying spongiotic annulus. A few small bundles of smooth muscle cells can be seen in the ventricularis of the basal part of the cusp. Small blood vessels are present only in the annulus and normally absent in the leaflet. Immunohistochemistry of the cuspal cell population. The observations are summarized in Table 2. The study shows: 1. Some expected findings: Factor VIII–positive endothelial cells on the cuspal surfaces; the diffuse presence of vimentin-positive, actin-negative fibroblasts in the three layers; a small number of actin-positive, vimentin-negative smooth muscle cells, the latter cells being essentially localized in the annulus and the ventricularis of the basal part of the cusps. 2. The presence in the annulus, ventricularis, and spongiosa of the cusps of many S100 protein-positive and actinnegative cells with a dendritic morphology, these features being characteristics of Lcs (Figure 1). In one case, a 13-year-old male donor who had been transplanted for dilated cardiomyopathy, the density of the Lcs was strikingly high, and they were also present in the fibrosa. 3. The infiltration of the cusps by a small to fair number of macrophages in four valves out of five. The macrophages were generally fusiform and had a quiescent aspect. Finally, the presence or absence of these different cells in the cusps as well as the cellular density did not appear to be associated with the age (except for case 1) or the clinical status of the donor. TEM study of the preservation state of cells and matrix in the cusps. The results of the study are presented in Table 3. Noncryopreserved valves. Samples of noncryopreserved valves gave rather fair pictures of fibroblasts collagen, elas-

38

GOFFIN ET AL. CRYOPRESERVED HOMOGRAFT VALVES

Cardiovasc Pathol Vol. 6, No. 1 January/February 1997:35–42

Figure 1. Anti-S100 positive Langerhans cells (arrows) in a nonimplanted cryopreserved aortic valve from a 14-year-old female (310).

tin, and GAGs. Endothelial cells were intact in four of five cases. Only a single macrophage was observed, except in one valve with (histologically) atheromatous lesions where many of these cells were present. The GAG study showed a pattern of small and medium size particles with regular distribution corresponding to collagen fibril periodicity. Larger GAGs were regularly localized between the fibrils. In some cases, the larger GAGs appeared to be clumped at the margin of the collagen bundles, and the smaller ones were less evident. Cryopreserved valves. Samples of cryopreserved valves with short-term cold ischemic time before freezing (three

cases, 20–30 hours) revealed nearly similar pictures as the nonpreserved controls. In a coronary artery disease case, the fibroblastic population of the valves dissected from the explanted heart was totally devitalized. In the group with mid-term cold ischemic time (5 cases, 30–50 hours) nearly the same patterns were observed. In one case, fairly preserved endothelial cells were still present (Figure 2), with only a few being disrupted. Some macrophages showed autolytic changes of the nucleus, the mitochondria, and membranes. In the series with long-term cold ischemic time (four cases, 50–62 hours) lytic lesions of macrophages

Table 3. Unimplanted Homograft Valves (TEM) Case Valve No. Type

Clinical Diagnosis Donor

Noncryopreserved 1 P Dilated CMP 2 P Asphyxia; long-term A Corticoids 3 A Dilated CMP Cryopreserved 4 P CAD 5 A CAD 6 P CAD 7 P CAD 8 A CAD 9 P Brain hemorrhage 10 P Brain tumor 11 A Polytrauma 12 A Dilated CMP 13 A Brain hemorrhage 14 A Digestive hemorrhage 15 A Heart infarct

Cold Ischemic Time (hours) 15.30 19.30

Endothelium

Fibroblast

Collagen

Elastin

GAGs

Fair Good Dark Fair

Good Good Good Fair

Good Fair (clumps) Good Good

36.30

Lost Lost* Lost Lost

25.40 27.15 29.25 33.00 39.05 43.30 46.25 48.30 50 .50 55.15 61.50

Lost Lost Intact Lost Lost† Fair Lost Lost Subtotal damage Subtotal damage Lost Lost

Fair Devitalized Good Fair Fair Fair; dark Good Fair Fair (blebs) Fair (clumped nuclei) Fair Fair (clumped nuclei)

Good Vesicles Fair Fair Good (vesicles) Fair Good Fair Fair Fair (vesicles) Fair Fair

Good Fair Good Good Fair Fair Fair Fair (clumps) Fair Good Fair Fair Fair (clumps) Fair (clumps) Fair Faint (clumps) Clumped Fair Clumped

Abbreviations: CAD 5 coronary artery disease, CMP 5 cardiomyopathy. * Endothelium well preserved in histology. † Not identified.

Fair

Fair (clumps)

Macrophages — — — Good to fair Good to fair

Good to fair Good to fair Bad Bad Bad

Cardiovasc Pathol Vol. 6, No. 1 January/February 1997:35–42

Figure 2. Cryopreserved valve with mid-term cold ischemic time (case 9: 43: 30 hours). Covering of the leaflet surface with fairly well-preserved endothelial cells is still present. (Lead citrate staining 344,000.)

GOFFIN ET AL. CRYOPRESERVED HOMOGRAFT VALVES

39

were more obvious, and some fibroblasts contained a condensed clumped nucleus. Numerous membrane-bound vesicles, suggesting cellular remnants, were interspersed between the collagenous fibrils (Figure 3A). The GAGs also were of less quality with evident clumping of the larger GAGs and faint staining or loss of smaller ones (Figure 3B). Biochemical control of GAGs. Treatment with chondroitinase ABC (substrate: chondroitin sulphate and dermatan sulphate) revealed a loss of the smaller GAGs, whereas chondroitinase AC treatment (substrate; chondroitin sulphate, dermatan sulphate and hyaluronic acid) showed a loss of small and larger GAGs; treatment with nitrous acid (substrate; heparan sulphate and heparin) had no effect. This indicates that the smaller cuprolinic blue–stained structures represented chondroitin and/or dermatan sulphate GAG– containing proteoglycans, whereas the larger ones contained hyaluronic acid. Moreover, heparan sulphate appeared not to be a major GAG present in the valves. Short-term implants. The pathologic findings are summarized in Table 4. Cases are listed according to the implantation duration. Analysis of the observations can be summarized as follows: 1. In a 3-hour pulmonary valve implant with early cuspal rupture (case 1), most endothelial cells had disappeared, whereas small hemorrhage and infiltration of polymorphonuclear leukocytes testified to an early inflammatory

Figure 3. Cryopreserved valve with long-term cold ischemic time (case 13; .50 hour): (a) Numerous membrane-bound vesicles, presumably remnants of disrupted cells, are intermingled with the collagenous stroma. (Lead citrate staining 363,000.) (b) A rather normal pattern of small, medium-size, and large GAGs is observed in the longitudinal section of the collagen. (Cuprolinic blue/tungstate staining 363,000.)

40

GOFFIN ET AL. CRYOPRESERVED HOMOGRAFT VALVES

Cardiovasc Pathol Vol. 6, No. 1 January/February 1997:35–42

Table 4. Pathology of Short-Term Implants (Listed According to Implantation Duration) Implantation Data

Healing Process

Age Patient (years)

Site

Duration

1

50

LVOT

3 hours

2

22

LVOT

8 days

3

22

RVOT

4

51/2

5

Case No.

Histopathology

Valve Function

Inflammatory Cells

Thrombotic Deposition

In Cusp (Fibroblasts)

On Cusp Surface‡

Residual Cellularity (%) Fibrocytes

Endothelium

Tissue failure† Normal

Neutrophils, hemorrhage Neutrophils

0

0

0

100

6 20

0

0

0

0

10 days

Technical failure

Rare macrophage

0*

0

50

0*

RVOT

10 days

Normal

1 (micro)

0

20

0

64

LVOT

10 days

Normal

6

60

LVOT

20 days

7

21

RVOT

21 days

Technical failure Normal

Neutrophils and mononuclear leukocytes rare immunobl Neutrophils and monuclear leukocytes 0

1 Annulus 11 Fibrosa and spongiosa 0

8§

35

LVOT

30 days

Technical failure

0 0

0

1 Annulus

0

20

0

0

0

10

0

11 (small/3 cusps) 0

0

1 1 commissure 11 2 cusps 1

0

0

50

0*

1 Spongiosa

* Immunohistochemical reaction of Lcs (anti-S100) and T-lymphocytes (anti-CD3) negative. † Thawing accident. ‡ Fibrous sheathing. § One cusp only.

2.

3.

4.

5.

reaction. The rupture was ascribed to a faulty handling of the graft during the thawing process. From postoperative day 10 on, all endothelial cells were destroyed. In an 8-day implant with normal function removed at autopsy (case 2), all cell types were lost. (This early devitalization might have resulted first from the fact that the donor died from carbon monoxide intoxication and second, that the graft had been decontaminated with Amphotericin, a very toxic antifungic drug). In cases 3 (day 10) and 8 (day 30), the only two valves submitted to immunohistochemistry, anti-S100–positive cells were no longer identified. Donor fibroblasts started to devitalize and vanish from postoperative day 10 leaving a residual cellularity varying between 50% and 0%. Signs of acute inflammatory reaction were precocious (case 1), but lasted only 6 2 weeks (less than 20 days). Mononuclear leukocytes (either unspecified, immunoblastic, or macrophagic) were observed only in the three 10-day implants (Figure 4). On the other hand, these cells were no longer present in the 20-, 21-, and 30-day implants. Microscopic small thrombosis was observed in two normal functioning valves, both implanted in the RVOT of a child (cases 4 and 7). In case 4, a pulmonary valve implanted in a 5 1⁄2-month-old child for 10 days and sampled at autopsy, small fresh thrombi were present in the valvular sinus and the ventricular surface of the cusps; no fibrous sheath

was present. In case 7 thrombosis was observed on a pulmonary valve 21 days after implantation in a 2-day-old baby: grossly visible small thrombi covering the aortic surface of the cusps were partly organized and coexisted in two cusps with a thin fibrous sheath, extending from the neighboring surface and containing few small vessels. Remarkably, the development of a fibrous sheath on the conduit wall and cuspal surfaces became microscopically apparent from day 20 onward (Figure 5) and might be considered as an early healing process originating from the host. 6. Intracuspal healing was expressed in many grafts from day 8 on by the penetration of young fibroblasts in the annulus (cases 2 and 5; Figure 4) and certain limited areas of the fibrosa (case 3) and the spongiosa (cases 3 and 8). 7. No obvious qualitative or quantitative differences in valvular changes were observed between either the type of homograft valve used (pulmonary versus aortic) or their site of implantation (right versus left VOT).

Comments As shown in the electron micrographs of cryopreserved unimplanted valves, short- or mid-term cold ischemia preceding cryopreservation does not alter significantly the quality of the valve constituents compared with noncryopreserved controls. Long-term cold ischemic time, however, seems to in-

Cardiovasc Pathol Vol. 6, No. 1 January/February 1997:35–42

GOFFIN ET AL. CRYOPRESERVED HOMOGRAFT VALVES

41

Figure 4. Fibrous annulus of an aortic sleeve homograft implanted as a miniroot in a 64year-old male. The patient died suddenly 10 days postoperation from rhythm disturbance, and the homograft was removed at autopsy (case 5). The largely devitalized fibrosa is infiltrated by both mononuclear leucocytes (mainly in the left part of the photomicrograph) and young fibroblastic/mesenchymal cells in the more superficial area (right part). (Hematoxylin-eosin-saffron 34.)

fluence graft quality, especially the ultrastructural integrity of macrophages and GAGs. Thus, despite the various stresses to which homograft valves are submitted during processing, neither the cold ischemic time nor the process of decontamination, cryopreservation, and thawing seem to alter dramatically the ultrastructure of the matrix constituents (collagen elastin and GAGs) of the cusps. It must be stressed that preservation of the ground substance and its GAGs not only guarantees a normal mechanical function to the fibrous network of the cusps, but also might play a role in the prevention of matrix mineralization by masking the sites where phosphate bonds initiate calcification (10,11).

Figure 5. Apical part of a homograft cusp (V) retracted and covered by the fibrous sheath (FS) of the recipient 21 days after implantation. RVOT reconstruction in a 2-day-old baby with pulmonary atresia and VSD (case 7). (Hematoxylin-eosin-saffron 310.)

The endothelium, however, is particularly sensitive and often damaged or lost before freezing and thawing. This observation, which is confirmed in the large samples by histology, is important because endothelial cells are immunogenic. As demonstrated in our immunohistochemical study, another immunogenic cell is present in the heart valve cusps as well, often in large numbers (designated as dendritic or Lc). One should expect that both the endothelial and Lcs would become early targets of rejection after implantation in patients. Our investigation on short-term explants confirms a similar finding (9,12). As mentioned in the introduction, these reports indicate that the homograft valves after cryopreservation (9,10,12) are still immunogenic and induce an early humoral

42

GOFFIN ET AL. CRYOPRESERVED HOMOGRAFT VALVES

immune response. It has been suggested that the reaction is clinically reversible without immunosuppression (8), and it might explain the fever seen after aortic valve replacement with a homograft (7). In our series not a single endothelial cell was observed on the surface of the homografts after 8 days implantation and no Lcs were identified (in the cusps of the two implants submitted for immunohistochemistry so far). On the other hand, the infiltration of T-lymphocytes did not represent in our series a significant component of the inflammatory reaction. The fibroblastic population of the cusps also seems to suffer after implantation and to start to devitalize after a few days. In our study, the estimated percentage of surviving matrix cells varies between 50% and 0% after 10 days of implantation. From this period onward many cells are macrophages or nonspecific mononuclear cells, both penetrating in the cusps in small to medium numbers. The real significance of this rarefaction of fibroblasts in the aftermath of the acute mainly humoral process of rejection is difficult to interpret: is it resulting from an ongoing low-grade immunogenic reaction, is it the end stage of a metabolic disturbance, or is it caused by both? Moreover, are the matrix and mononuclear cells from donor or recipient origin? A comparison of our observations on cryopreserved valves with those made in acellular antibiotics preserved homografts (13) and a full pathology study of mid- to long-term cryopreserved valve implants, including DNA fingerprinting specifying the origin of the residual cells, are in progress. It is expected that they will give an insight into the fate of implanted cryopreserved viable homograft valves and indicate (i) whether they will keep a well-preserved collagenous structure and last longer than other biologic valves and (ii) whether they will resist to wear and tear longer than their antibiotics-treated devitalized counterparts, despite their having been submitted to a stronger humoral rejection after implantation and the evidence of early and progressive fibroblast rarefaction. Although this explant study did not show any obvious differences in pathologic changes neither between pulmonary and aortic homografts nor between right- and left-sided valves, one might expect mid- and long-term implants to behave differently, at least at the quantitative level. The knowledge that accelerated mineralization of bioprosthetic valves implanted in children is associated with an increased calcium turnover also suggests that this phenomenon also occurs early in devitalized homograft valves implanted in the same age group. All these presumed qualitative differences are currently observed in a study to be published.

Conclusions 1. The finding in the cusps of unimplanted cryopreserved valves of numbers of residual endothelial cells and of an important population of dendritic cells indicates that the

Cardiovasc Pathol Vol. 6, No. 1 January/February 1997:35–42

cusps are still immunogenic at implantation. Furthermore, the early postimplantation disappearance of these two types of cells, together with the scarcity of a specific inflammatory reaction, offers indirect evidence that an early humoral rejection takes place soon after implantation. 2. The reason why large proportions of the fairly well-preserved fibroblastic population of the cryopreserved homograft valves devitalize early and progressively after implantation remains unclear, but it is likely that both low-grade lasting rejection and eventual metabolic cell disturbances play a role. 3. Finally, the good to fair preservation before implantation of all the constituents of the valvular matrix and the ground substance (collagen and GAGs) might turn out to be favorable to the long-term fate of the implants. But this needs to be proven by in vitro mechanical testing and thorough ultrastructural studies of mid- and long-term implants.

References 1. Kirklin JK, Naptel DC, Novick W, et al. Longterm function of cryopreserved aortic valve homografts: a 10 year study. J Thorac Cardiovasc Surg 1993;106:154–166. 2. Doty DB, Michielon G, Wang N-D, Cain AS, Millar RC. Replacement of the aortic valve with cryopreserved aortic allograft. Ann Thorac Surg 1993;56:228–236. 3. O’Brien MF, Stafford EG, Gardner MAH, Pohlner PG, McGiffin DC, Kirklin JW. A comparison of aortic valve replacement with viable cryopreserved and fresh allograft valves, with a note on chromosomal studies. J Thorac Cardiovasc Surg 1987;94:812–823. 4. O’Brien MF, Stafford EG, Gardner MAH. Allograft aortic valve replacement: long-term follow-up. Ann Thorac Surg 1995;60:S65–S70. 5. Shapira M, Fonger JD, Reardon K, Shemin RJ. Unexplained fever after aortic valve replacement with cryopreserved allografts. Ann Thorac Surg 1995;60:S151–S155. 6. Fischlein T, Schütz A, Haushofer M, et al. Immunologic reaction and viability of cryopreserved homografts. Ann Thorac Surg 1995;60: S122–S126 7. Smith JD, Ogino H, Hunt D, Laylor RM, Rose ML, Yacoub MH. Humoral Immune Response to Human Aortic Valve Homografts. Ann Thorac Surg 1995;60:S127–S130. 8. Mitchell RN, Jonas RA, Schoen FJ. Structure-function correlations in cryopreserved allograft cardiac valves. Ann Thorac Surg 1995;60:S108– S113. 9. Van Kuppevelt THMSM, Cremers FPM, Domen JGW, Kuyper CMA. Staining of proteoglycans in mouse lung alveoli. II. Characterization of Cuprolinic blue-positive, anionic sites. Histochem J 1984;16:671–686. 10. Scott JE. Proteoglycan-collagen interaction. In: Everard D, Whelen J, eds. Function of Proteoglycans. Ciba Foundation Symposium 124. Chichester: Wiley, 1986;104–124. 11. Poole AR, Webber C, Pidoux I. Localization of dermatan sulfate proteoglycan (DS-PG II) in cartilage and the presence of immunologically related species in other tissue. J Histochem Cytochem 1986;34:619–625. 12. Hoekstra F, Knoop C, Aghai Z, et al. Stimulation of immune-competent cells in vitro by human cardiac valve-derived endothelial cells. Ann Thorac Surg 1995;60:131–134. 13. Goffin YA, Black MM, Lawford PV. The stability and performance of bioprosthetic heart valves. In: Williams DF, ed. Current Perspectives on Implantable Devices. London: JAI Press, 1990:85–92.