Donor-Host Involvement in Immature Rat Testis Xenografting into Nude Mouse Hosts

Abstract

Immature testicular tissue of a wide variety of mammalian species continues growth and maturation when ectopically grafted under the dorsal skin of adult nude mouse recipients. Tissues from most donor species fully mature, exhibiting complete spermatogenesis within months. The connection to the recipient's vascular system is mandatory for graft development, and failure of vascularization leads to necrosis in the grafted tissue. In the present study, we analyze to what extent 1) the xenografted immature donor tissue and 2) the recipient's cells and tissues contribute to the functional recovery of a “testicular xenograft.” We address whether recipient cells migrate into the testicular parenchyma and whether the circulatory connection between the donor testicular tissue and the recipient is established by ingrowing host or outgrowing donor blood vessels. Although this issue has been repeatedly discussed in previous xenografting studies, so far it has not been possible to unequivocally distinguish between donor and recipient tissues and thus to identify the mechanisms by which the circulatory connection is established. To facilitate the distinction of donor and recipient tissues, herein we used immature green fluorescent protein-positive rat testes as donor tissues and adult nude mice as graft recipients. At the time of graft recovery, donor tissues could be easily identified by the GFP expression in these tissues, allowing us to distinguish donor- and recipient-derived blood vessels. We conclude that the circulatory connection between graft and host is established by a combination of outgrowing small capillaries from the donor tissue and formation of larger vessels by the host, which connect the graft to subcutaneous blood vessels.

The circulatory connection between donor and host in testis xenografting is established by outgrowing donor and ingrowing host blood vessels.

INTRODUCTION

Xenografting of testicular tissues has been explored as an approach to study the physiology of the immature testis [1–5] and as a potential tool for fertility and germline preservation in pediatric patients with cancer [6–9]. It might also offer an approach for conservation of rare or endangered species [10–12] and domestic breeds [13–16]. In most cases, small pieces of dissected testicular tissue have been ectopically (e.g., under the dorsal skin) grafted into nude mouse hosts. In some experiments, cell suspensions derived from immature donor testes have been transplanted into nude mouse hosts either directly or after a transient period of in vitro culture [17–19]. While potential clinical applications of testis xenografting have already been discussed, the basic processes that influence the outcome of testicular xenografting have yet to be elucidated. Only xenografting of immature donor testicular tissue has resulted in successful progression of grafts to full spermatogenesis. All attempts to xenograft adult testicular tissue have failed, as these grafts did not contain differentiating male germ cells [6, 20, 21]. The failure of adult testicular tissue to establish functional xenografts has been attributed to the high demand for oxygen in adult testicular tissue with ongoing spermatogenesis. The high metabolic activity of seminiferous epithelium leads to a prolonged and fatal period of hypoxia between dissection and reconnection of the xenografted tissue to the circulatory system of the host [6, 21]. Thus, after xenografting of fully adult testicular tissues, only degenerated and fully hyalinized testicular tissue has been recovered. In contrast, immature testicular tissue of most mammalian species xenografted into nude mouse recipients survives and grows. Depending on the species, it takes several weeks or months after grafting until at least some seminiferous tubules exhibit full spermatogenesis and steroidogenesis [2, 21]. Xenografted immature testicular tissue establishes a connection to the circulatory system of the host mouse, thus securing an adequate supply of oxygen and nutrients and providing a route for the release of androgens. It can be speculated that the rather inactive immature testicular tissue has a low demand for oxygen compared with adult tissue, leading to less severe hypoxia in the transition period before establishment of the circulatory connection. It is unknown how the blood supply between grafted testicular tissue and host is being established. Blood vessels from the grafted tissue could outgrow to connect to blood vessels of the host, or capillaries from the host could grow into the implanted donor tissue. It is also unresolved whether the host contributes any cells or components to the recovering and differentiating donor tissue. Each testicular xenograft is enclosed by a sturdy capsule, but the origin of cells in this capsule has not been explored to date. Other cells such as macrophages or mensenchymal precursors could also migrate into the grafted tissue and thus provide precursors for interstitial cells.

The aim of this study was to explore the contribution of cells from donor and host in establishing fully functional xenografts. We used rats (SD-Tg[CAG-EGFP]CZ-004Osb rats) that ubiquitously express green fluorescent protein (GFP) as testis tissue donors and nude mice (strain Nu/Nu) as hosts. The questions addressed in this study were 1) whether the blood supply to grafted testis tissue is established by ingrowing donor or outgrowing host blood vessels and 2) whether host cells migrate into the donor testicular tissues and take part in formation of intraparenchymal blood vessels or interstitial tissues.

MATERIALS AND METHODS

Animals

A colony of rats expressing GFP (SD-Tg[CAG-EGFP]CZ-004Osb rats) [22, 23] was established using breeders that were kindly provided by Dr. M. Okabe, Osaka University, Osaka, Japan. Sixty-two newborn GFP rat pups were obtained from our colony as testis tissue donors, and 70 GFP rat pups at age 7 days postpartum were used as testis donors for the preparation of testicular single-cell suspensions. We chose the GFP-positive rat as testis tissue donor because preliminary experiments demonstrated that the immature GFP-positive rat testis tissue shows consistent and reproducible graft development and survival and that this grafting model yields much more reliable results than mouse-to-mouse grafting trials (e.g., LacZ-positive testis into a host with a genetically identical ROSA26 background).

Male immunodeficient nude mice (strain Nu/Nu) at age 5–7 wk were obtained from Charles River Laboratories (Wilmington, MA) or from Janvier Europe (Le Genese St. Isle, France) and were maintained in our rodent housing facilities until they reached age 12 wk.

Most animals were maintained at the rodent housing facilities of the University of Pittsburgh School of Medicine and the Magee-Womens Research Institute under 12L:12D, with pelleted food and water available ad libitum. The animals used in the in vivo monitoring of graft development were maintained at the central animal facilities of the School of Medicine of the University of Münster, Germany, under the same light conditions. Animal husbandry and all experimental procedures involving the animals were performed in compliance with the University of Pittsburgh and the Magee-Womens Research Institute Guidelines for the Care and Use of Laboratory Animals and in accord with the German Federal Law on the Care and Use of Laboratory Animals (license G24/2006).

Tissue Preparation and Cell Culture

For testis tissue grafting, newborn rat pups were killed by decapitation, and the testes were excised. The epididymis was removed, and each testis was longitudinally cut in two halves. During preparation and before grafting, testis tissue was maintained for less than 2 h in ice-cold cell culture medium (L-15 Leibovitz; Mediatech Inc., Herndon, VA) with added nonessential amino acids and antibiotics and 5% fetal calf serum.

For cell culture grafting, rat pups at age 7 days were killed by decapitation, and the testes were removed. Single-cell suspensions of testicular cells were prepared by sequential enzymatic digestion [24]. In brief, testes were decapsulated, and the seminiferous tubules were first digested with 1 mg/ml of collagenase I (C-2674; Sigma, St. Louis, MO) and 5 μg/ml of DNase (15 U/ml, 104132; Roche Applied Science, Indianapolis, IN) in Dulbecco Minimum Essential Medium (DMEM) (4.5 g/ml of glucose). Isolation of seminiferous tubular fragments from interstitial cells was achieved by repeated sedimentation at unit gravity. In a second digestion step, tubule fragments were incubated with 1 mg/ml of collagenase I and 5 μg/ml of DNase in combination with 1 mg/ml of hyaluronidase (H-3506; Sigma) until a single-cell suspension was achieved, which was washed and resuspended in low-glucose DMEM (1 g/L of glucose) supplemented with nonessential amino acids and antibiotics. Total cell numbers were assessed using a bright-line hematocytometer (3100; Hausser Scientific, Horsham, PA).

In each well of a 24-well cell culture dish, 1 × 10⁶ rat cells were plated on 250 μl of reconstituted extracellular matrix (ECM) gel (354234, Matrigel; BD Biosciences, Bedford, MA) diluted 1:1 with culture medium (DMEM containing 1 g/ml of glucose, antibiotics, and nonessential amino acids). The Matrigel coating was prepared before plating the cells to allow the stated number of cells, suspended in 1.5 ml of culture medium, to be seeded on top of the Matrigel coating. The Matrigel coating was approximately 300 μm thick in the center of each well and thicker toward the edges. Cells were cultured for 9 days in an incubator at 35°C in an atmosphere containing 5% CO₂ and saturated humidity.

Grafting

At age 12 wk, recipient nude mice were castrated under anesthesia (ketamine [80 mg/kg of body weight] and xylazine [6 mg/kg of body weight] i.p. in saline) through scrotal incisions, and eight halves of newborn GFP rat testes per recipient were placed under the dorsal skin using cancer implant needles (G13; Popper and Sons, Staunton, VA). Alternatively, six injections of 250 μl of ECM containing the cellular aggregates after 9 days of culture were placed under the dorsal skin of each recipient using G18 injection needles and tuberculin syringes (Fig. 1). As the cultured cell aggregates are injected in a viscous suspension in ECM, single injections can separate into two morphologically distinguishable structures after grafting, thus leading to a higher number of recovered grafts than the number of ECM injections originally applied to the recipient.

PLATE I. Figures 1 and 2. FIG. 1. Schema demonstrating the grafting procedure. GFP-positive rat pups can be distinguished from GFP-negative littermates by UV illumination (left). GFP-positive testes are excised (I), transversely cut in halves (II), and grafted under the dorsal skin using G13 transfer needles (III).FIG. 2. In vivo fluorescence reflectance image of a nude mouse carrying GFP-positive rat testicular grafts imaged at 3–10 days after graft implantation. The grafted GFP-positive tissues can be clearly detected under the dorsal skin of the host (arrowheads). Note the increasing size of the grafts over time.

In Vivo Monitoring of Graft Development

In addition to the hosts grafted with GFP-positive immature rat testes for the histological evaluation already described, six additional recipients were castrated and grafted with six halves of immature GFP-positive rat testes. On Days 3, 5, 7, and 10 after grafting, the animals were anesthetized using ketamine-xylazine as already described, and the grafts were visualized and images recorded (In-vivo Imaging Station FX Pro; Kodak, Stuttgart, Germany). At Day 14, the animals were killed under deep anesthesia, the dorsal skin was partially removed to expose the grafts in situ, and micrographs of the grafts and their blood vessels were obtained using a dissecting scope (SZX10; Olympus, Melville, NY) with a camera (SC20; Olympus) and imaging software (Cell^A; Olympus) and an inverted microscope (Axiovert 200; Zeiss, Welwyn Garden City, England) with a monochrome camera (Spot Insight 2, Diagnostic Instruments, Inc., Sterling Heights, MI) using the manufacturer's imaging software.

Sample Retrieval and Preparation for Histological Evaluation

Most grafts were harvested at 4 and 8 wk after implantation. At 4 and 8 wk, maturation of seminiferous epithelium can already be assessed by the presence or absence of differentiating and meiotic germ cells, but the graft is not yet producing sperm, which could lead to deteriorating conditions within the grafts because of the lack of an efferent system and thus a large amount of disintegrating cells (sperm) in the tubular lumen. Thus, these time points were chosen to allow for maximum graft survival and tissue integrity. One hour before killing the recipients, they were injected i.p. with a 1% solution of 5-bromo-2′-deoxyuridine (BrdU) (B9285; Sigma) in 0.9% sterile saline (100 mg/kg of body weight). For graft retrieval, nude mouse recipients were killed by exsanguination under deep anesthesia. The dorsal skin of each mouse was removed, and grafts were located on the interior surface of the skin. In all recipient animals, body weight, seminal vesicle weight (SVW) (as an indirect indicator for androgenic activity of the grafted cells and tissues in the castrated recipients), and the number of encountered grafts were determined. The weight of each graft was also recorded. Grafts were excised and fixed in Bouin fixative overnight at room temperature, followed by storage in 70% ethanol. Some grafts were fixed in 2% paraformaldehyde solution in phosphate buffer for 4 h and then transferred to 30% sucrose solution (overnight). Those grafts were then frozen in intermedium (2-methylbutane, UN1265; Acros, Geel, Belgium), which was cooled by liquid nitrogen. Samples were then embedded in freeze-embedding medium (Neg-50; Richard Allan Scientific, Inc., Kalamazoo, MI).

Sectioning and Immunohistochemistry

Freeze-embedded samples were sectioned to 10 μm on a cryotome (Microm Cryo-Star HM 560 MV; MICROM International GmbH, Walldorf, Germany). Bouin-fixed samples were transferred and stored in 70% ethanol and embedded in paraffin. Sections of 5 μm were prepared on a motorized mictrotome (HM 360; MICROM International GmbH). Paraffin sections were immunohistochemically stained using primary antibodies against GFP (rabbit polyclonal anti-GFP, 290; Abcam, Cambridge, MA), against smooth muscle actin (SMA) (mouse monoclonal anti-SMA, A2547; Sigma), and against BrdU (monoclonal anti-BrdU clone BU33, B2531; Sigma). Cryosections were only stained to detect SMA. Primary antibodies were detected using secondary antibodies directly or indirectly labeled with green fluorophores (Goat Anti-Rabbit Alexa 488 to label GFP) [A-11008]; Invitrogen, Carlsbad, CA), red fluorophores (Goat Anti-Mouse Alexa 546 to label SMA [A-11003]; Invitrogen), or biotin (for detection of BrdU). Fluorescence-labeled slides were then coverslipped with mounting medium (Vectashield; Vector Laboratories, Burlingame, CA) containing 4′,6-diamidin-2′-phenylindol-dihydrochloride (DAPI) to obtain a nuclear counterstain. BrdU detection was accomplished by incubating the slides in a third step with a streptavidin-conjugated horseradish peroxidase (S5512; Sigma), 3,3′-diaminobenzidine as chromogen (D4168; Sigma), and hematoxylin as a counterstain, followed by permanent coverslipping.

Histological Sample Evaluation and Image Acquisition

Samples were analyzed microscopically (BX61; Olympus) using fluorescent illumination equipment with an attached camera (Retiga 4000R; QImaging, Burnaby, BC). All images were acquired digitally using imaging software (Northern Eclipse 7.0; Empix Imaging Inc., Cheektowaga, NY) and were processed using a software program (Photoshop 6.0; Adobe Systems Inc., San Jose, CA).

One slide carrying representative cross-sections of each graft was evaluated by one observer (J.E.) in a blinded manner. The parameters determined for each graft included general graft composition (fibrosis, fatty tissue, leukocyte infiltration, and the presence of seminiferous and rete testis tubules) and the most advanced germ cell type present in each tubule. One section through the center of each graft was randomly chosen. Within this section, every round cross-section of all seminiferous tubules was evaluated.

Data Evaluation

Analysis was performed using statistical software (SigmaStat 3.1.1; Systat Inc., Point Richmond, CA). Average body weight, SVW, and graft weight were compared between the two different time points (4 vs. 8 wk). Independent data analysis was performed for grafted tissue and grafted cells. Average body weight, SVW, and graft weight from all recipients that had received tissue grafts were compared with the same parameters from the group that had received cell culture grafts (regardless of the retrieval time point). The statistical analysis was performed using Student t-test and/or Mann-Whitney rank sum test. Histological data derived from xenografts were collected as relative data per graft, including the relative number (percentage) of seminiferous tubules containing 1) no germ cells (Sertoli cell only [SCO]), 2) spermatogonia, 3) pachytene spermatocytes, or 4) spermatids as the most advanced germ cell type. Differences were considered statistically significant at P < 0.05.

RESULTS

Recipient Status

The mice that received testis tissue grafts had average body weights of 34.22 (±3.09) g at 4 wk and 33.86 (±3.24) g at 8 wk (Table 1); the difference between the two time points was not statistically significant. The average body weights of testis cell recipients differed at 4 wk (27.00 [±2.24] g) vs. 8 wk (31.08 [±3.68] g) (P = 0.037). Also, the average body weight of all testis tissue recipients (34.06 [±2.86] g) differed from the average body weight of testis cell recipients (29.15 [±4.00] g) (P < 0.001).

TABLE 1.

Anatomy of recipients and grafts.

The testis tissue recipients had average SVWs of 205.38 (±34.64) mg at 4 wk and 203.29 (±92.13) mg at 8 wk (Table 1). Testis cell recipients had average SVWs of 26.00 (±6.44) mg at 4 wk and 36.75 (±27.46) mg at 8 wk. Within each group, the SVW difference between the two time points did not differ significantly. However, the average SVW of all testis tissue recipients (206.18 [±61.16] mg) differed from the average SVW of all testis cell recipients (31.92 [±26.57] mg) (P < 0.001).

Graft Status

Growth, anatomy, and macroscopic appearance.

In the in vivo monitoring of GFP-positive grafts, clear growth of grafts occurred during the first 10 days after grafting (Fig. 2). The largest diameter of a graft was 5.5 mm at 3 days after implantation, while the largest diameter was 8.0 mm at 10 days after grafting.

Of 136 testis tissue grafts implanted for histological analysis, 90 grafts were recovered (51/80 [63.8%] after 4 wk and 39/56 [69.6%] after 8 wk) (Table 1). Of 102 cell injections, 101 grafts were recovered (24/30 [80%] after 4 wk and 77/72 [107%] after 8 wk). Testis tissue grafts had average weights of 154.59 (±26.98) mg at 4 wk and 128.07 (±52.45) mg at 8 wk. Testis cell grafts weighed 15.34 (±4.10) mg at 4 wk and 19.28 (±9.93) mg at 8 wk. Weight differences within each group between the two time points were not statistically significant. However, the total average graft weight for all testis tissue grafts (143.67 [±48.54] mg) differed from the total average weight of all testis cell grafts (18.12 [±8.68] mg) (P < 0.001).

The additional grafts that were removed on Day 15 had an average weight of 131.61 mg (range, 4–385 mg). The grafts showed a fine network of blood-filled capillaries (Fig. 3, a–c). Large vessels were observed connecting the grafts to subcutaneous blood vessels of the host (Fig. 3a). The grafts were clearly GFP positive, whereas the surrounding tissues and the supplying larger blood vessels were GFP negative (Fig. 3, c and d).

FIG. 3.

Images of xenografts derived from immature GFP-positive rat testes showing the grafts still attached to the partly removed dorsal skin of the recipient. a) Two grafts (arrowheads) show small blood vessels in capsule and parenchyma. A subcutaneous blood vessel (scv) can be seen directly connected (arrow) to a vessel supplying one of the grafts (gsv). b) Three grafts (asterisk) show fine blood vessels. Note the bigger subcutaneous vessels (arrowhead). c and d) GFP-positive grafts (arrowheads) can be seen lying on the GFP-negative background of the dorsal skin of the hosts. Blood vessels are clearly blood filled (arrows). The large graft-supplying blood vessel (d, arrow) appears GFP negative and is thus most likely of host origin.

Cellular composition.

In testis tissue grafts at the time of graft recovery, no seminiferous tubules with spermatids could be detected. Seventy of 90 grafts (77.8%) showed seminiferous tubules containing pachytene spermatocytes, and eight of 90 grafts (8.9%) contained spermatogonia as the most advanced germ cell type. One of 90 grafts (1.1%) showed seminiferous tubules without any germ cells (SCO), and 11 of 90 grafts (12.2%) did not contain distinguishable seminiferous tubules at all (Table 2). In addition, in grafts retrieved from three of 10 recipients that had received testis tissue grafts harvested after 4 wk, significant leucocyte infiltration was detected in the grafted testicular tissue, leading to complete loss of structure of seminiferous epithelium. Minor leucocyte infiltration could also be detected in most grafts.

TABLE 2.

Cellular status of the seminiferous epithelium in the grafts.*

In testis cell grafts, no seminiferous tubules with spermatids or pachytene spermatocytes could be detected. Four of 97 grafts (4.1%) contained spermatogonia as the most advanced germ cell type. Fifty-nine of 97 grafts (60.8%) showed seminiferous tubules without any germ cells (SCO), and 34 of 97 grafts (35.1%) did not contain distinguishable seminiferous tubules at all. Some leucocyte infiltration was also detected in most testis cell grafts.

GFP-positive cells and tissue.

Testis tissue grafts recovered at both 4 and 8 wk were always encapsulated by a two-layered capsule (Fig. 4a). The inner layer of this capsule was single layered and consisted of GFP-positive cells; the outer layer was multilayered and consisted almost exclusively of GFP-negative cells. Occasionally, GFP-positive single cells could be observed within the GFP-negative layer (Fig. 5b). The testicular parenchyma showed mostly fully developed tubules with a polarized GFP-positive seminiferous epithelium, a lumen, spermatogenic progression up to the pachytene spermatocyte stage, and a fully differentiated SMA-positive basement membrane (Figs. 4a and 5a). The cells in the testicular interstitium were GFP positive (Figs. 4a and 5a). Small blood vessels were located in the interstitium (Figs. 4a and 5c). The endothelia of those interstitial blood vessels were always both GFP positive and SMA positive. In the GFP-negative layer of the capsule, larger blood vessels could be observed. Some of these blood vessels had GFP-negative endothelia, whereas others had GFP-positive endothelia (Fig. 4b). In some samples, GFP-positive blood vessels were observed that were clearly protruding from the inner GFP-positive tissues out into the GFP-negative areas of the capsule (Figs. 4c and 5d). Erythrocytes could be observed in most of these blood vessels (Fig. 5d), indicating that these blood vessels are functional.

FIG. 4.

Fluorescent micrographs showing cryosections of GFP rat testis tissue grafts retrieved at 4 wk after implantation. GFP (green) is visualized by native fluorescence, and SMA (red) is visualized by immunohistochemistry. Nuclei are labeled with DAPI (blue). a) Note the GFP-positive inner capsule (arrow), the GFP-negative outer capsule (arrowhead), the blood vessels in the interstitium (double arrow), and the seminiferous tubules with ongoing spermatogenesis and lumen (asterisk). b) Note the GFP-positive (arrow) and GFP-negative (arrowhead) blood vessels in the GFP-negative part of the capsule. c) A GFP-positive blood vessel can been seen protruding into the GFP-negative part of the capsule (arrow).

FIG. 5.

Fluorescent micrographs showing paraffin sections of GFP rat testis tissue at 8 wk (a–c) and 4 wk (d) after implantation. GFP (green) and SMA (red) are visualized by immunohistochemistry; nuclei are labeled with DAPI (blue). a) Seminiferous tubules are surrounded by an SMA-positive basement membrane (arrow). Note the GFP-positive interstitium (asterisk). b) Few GFP-positive cells can be detected in the otherwise GFP-negative part of the capsule (arrows). c) Numerous small blood vessels are present in the interstitium (arrows). The inset in c (same magnification as panel c) shows a technical control for the immunohistochemical staining (omission of the primary antibodies). d) A GFP-positive blood vessel can be observed within subcutaneous muscles of the recipient mouse (arrow). Nonnucleated erythrocytes (cells appearing green with no nuclear DAPI label within the vessel lumen) can be detected within this blood vessel, underlining the functionality of this particular vessel.

In testis cell grafts, an inner GFP-positive capsule was also observed (Fig. 6a). Seminiferous tubules could be observed, and some of those showed a polarized seminiferous epithelium and an SMA-positive basement membrane. No lumen was observed within the seminiferous tubules.

FIG. 6.

Fluorescent micrographs showing paraffin sections of GFP rat cell grafts at 4 wk after implantation. GFP (green) and SMA (red) are visualized by immunofluorescence. Nuclei are labeled with DAPI (blue). The bar shown in c also applies to a and b. a) The graft is surrounded by a GFP-positive (arrow) and GFP-negative (arrowhead) capsule. Note the GFP-positive and GFP-negative cells in the interstitium (asterisk) and the poorly differentiated seminiferous tubules. b) Some grafts show a limited number of GFP-positive cells in the interstitium (asterisk), and only cord-like structures and tubular shadows can be observed. In some grafts, no seminiferous cords or tubules can be observed, although a few GFP-positive cells and tissues can still be found in the tissues (arrows). c) Few GFP-positive tubules can be observed within this graft (arrows), between a series of tubular shadows.

The interstitium was composed of both GFP-positive and GFP-negative cells (Fig. 6a). In many cases, tubulogenesis appeared incomplete, and these structures were referred to as cord-like structures (Fig. 6b). In other grafts, few GFP-positive tubules or cells could be distinguished within predominantly GFP-negative tissues (Fig. 6c).

BrdU.

BrdU had been injected i.p. into some of the recipient mice 1 h before tissue removal to assess whether blood from the recipient quickly circulated into the grafted tissues and thus accomplished a BrdU label in proliferating cells in the graft (BrdU is incorporated into newly assembled DNA during S-phase) within this limited time. BrdU was immunohistochemically detected in many cells of seminiferous epithelium and interstitium in four randomly chosen grafts retrieved from three BrdU-treated hosts (Supplemental Fig. S1 available at www.biolreprod.org), indicating that blood from the host had in fact carried BrdU into the grafted tissues and proving the efficiency of the circulatory connection between host and graft.

DISCUSSION

In the present study, we implanted immature testicular tissues and cultured testicular cells from rats ubiquitously expressing GFP into (GFP negative) nude mice. This allowed us to unequivocally identify the GFP-positive rat tissues (including blood vessels) in the nude mice weeks after implantation of these tissues. We used both regular testis tissue grafts and testicular cells cultured for 9 days in ECM gel to assess the involvement of donor and host cells and tissues in both settings. Whereas the focus in testis tissue grafting was on the revascularization of tissue, the focus in the cell grafts was primarily on the involvement of host cells in structural reorganization of seminiferous epithelium and reformation of interstitial cell cohorts. The anatomical data derived from the testis tissue recipients show that the host's body weight did not differ significantly at 15 days or at 4 and 8 wk after graft implantation, that the SVW weight was in the low normal range, and that the graft weights were not significantly different at any of the two time points. In the hosts that had received cells grafts, the host's body weight was higher at 8 wk than at 4 wk, and SVW was barely above castrated levels. This indicates that androgen production from grafts constituted from testicular cell suspensions was low. This finding is supported when data from all testis tissue recipients are pooled and compared with the pooled data of testis cell recipients. Recipients of cells have significantly lower body weight and much lower SVW, indicating that tissue grafts (but not grafts constituted from cell suspensions) provide androgen substitution to physiological levels.

Regarding the progression of spermatogenesis in the grafts, grafts from tissue demonstrate much better outcome than grafts from reconstituted cell suspensions. These findings are in accord with our earlier study [17], which showed that xenografted immature rat testis cells after 9 days of in vitro culture can reassemble seminiferous tubules with a polarized epithelium in the host but that few undifferentiated germ cell can be found in these grafts and no germ cell differentiation can be observed. In contrast, similar investigations using porcine cell suspensions showed full spermatogenesis [25]. It is unknown whether the absence of germ cells in rat grafts (but not in porcine grafts) is related to a species-specific difference or to methodological differences in preparing the cell suspension.

Our study shows that formation of the circulatory connection between the xenografted tissue and the host is initiated by outgrowing blood vessels from the grafted tissue, as GFP-positive (rat origin) blood capillaries were found outside of the GFP-positive core of the tissue grafts in the surrounding GFP-negative (mouse origin) connective tissue and muscle, but no ingrowing blood vessels from the host into the grafted tissue could be observed within the grafted testicular parenchyma. Nevertheless, the major blood vessels carrying blood to and from the graft and connecting the graft's blood vessels to larger subcutaneous blood vessels of the host are actually formed by the host.

The capsule enclosing the grafted tissue was composed of a GFP-positive inner layer and a GFP-negative outer layer. Because the newborn testis was not decapsulated but only cut into two halves before xenografting, the inner GFP-positive layer likely originates from the newborn rat testicular capsule. The GFP-negative outer capsule is provided by the mouse host. In many grafts, GFP-positive single cells were observed in the surrounding GFP-negative tissues, demonstrating migration of cells from the graft into the host. However, no GFP-negative cells were encountered within GFP-positive grafted tissue. Although it is difficult to identify GFP-negative cells within the GFP-positive interstitium, this indicates that few host cells, if any, migrated into the graft.

In contrast, when immature rat testis cells are xenografted after 9 days of in vitro culture, a limited reassembly of seminiferous tubules can be observed. However, even in grafts in which a polarized seminiferous epithelium is reestablished and peritubular cells can be detected by SMA expression, the interstitium is only partially made up of both GFP-positive and GFP-negative cells in interstitial-like spaces in xenografts, indicating migration of host cells into the grafts. In some cases, infiltration of GFP-negative cells also included the tubular compartment of the grafts, leading to complete loss of GFP-positive cells even within the former tubular structure and a loss of the functional seminiferous epithelium. In these testis cell grafts, GFP-positive blood vessels could only be seen between the few distinguishable seminiferous tubules, but no outgrowing vessel could be observed.

Xenografting of testicular tissues into nude mice was described in 2002 as an approach to explore testicular physiology and development [26]. Subsequent studies [2, 5, 7, 8, 27] have shown the efficiency of the method for derivation of sperm from sexually immature donors in a variety of rodent and primate species. Additional studies [10, 12, 13, 15, 16, 21, 28–30] have successfully established spermatogenesis in xenografted immature testicular tissues from domestic and companion species. Also, the use of xenografted nonhuman primate testicular tissue in studies addressing gonadotoxicity issues has been demonstrated [3], and xenografting has been used in investigations focusing on testicular aging [31] and as an assay for leukemic infiltration of testicular tissue [32, 33]. Xenografting of testicular tissues has therefore been proposed as a tool for fertility preservation in endangered species and in pediatric patients with cancer [31].

Little information is available on the performance of human testicular tissues as xenografts. Several groups have showed limited survival of adult and prepubertal human testicular tissue xenografts [6, 9, 20, 34].

Most of the aforementioned research revealed that immature testis tissue has much better developmental potential for xenografting than adult testis tissue. Some findings suggest that poor performance of adult testis tissue grafts is due to high oxygen and nutrient demand of adult tissue, leading to hypoxic conditions and thus to degeneration of the grafted tissues. Immature testis tissue has lower oxygen consumption and better odds of surviving the period of hypoxia after implantation. To establish a good supply of oxygen and nutrients, the grafted tissue must establish a blood vessel connection to the host's circulatory system quickly and efficiently. Previous studies demonstrated that xenografted porcine ovarian tissues [35], human ovarian tissues [36], and testicular tissues [17] show abundant revascularization. Furthermore, it has been suggested that maturation of some but not all seminiferous tubules in xenografts might be caused by insufficient nutrient and oxygen supply in some areas of the grafted tissue [5]. Overall, key questions in the past have been (1) whether grafted tissue establishes the circulatory connection with the host by attracting ingrowing blood vessels from the host or by initiating outgrowth of donor-originating blood vessels into the surrounding host tissues to obtain an anastomosis to the host's blood vessels and (2) whether the circulatory connection between graft and host is actually functional, supplying the grafts with blood from the host. In the present study, BrdU (which has a short half-life in the circulation of injected hosts [37, 38]), after injection into the host 1 h before graft removal, was found in proliferating cells within the graft's seminiferous epithelium and interstitium, clearly indicating a fully functional blood supply from host to graft.

In conclusion, we demonstrate herein that testis tissue xenografts derived from immature GFP-positive rat testes develop well as xenografts in nude mouse recipients. Furthermore, the circulatory connection between grafted tissue and recipient is established by a combination of outgrowing small capillaries from grafted tissue into surrounding host tissues and formation of larger blood vessels by the host, which then connect the blood vessel network of the graft to larger subcutaneous blood vessels of the host.