Heterotopic uterine transplantation by vascular anastomosis in the

Heterotopic uterine transplantation by vascular anastomosis in themouse

R Racho El-Akouri, G Kurlberg1, G Dindelegan1, J Mölne2,A Wallin and M BrännströmDepartment of Obstetrics and Gynecology, The Sahlgrenska Academy at Göteborg University, Göteborg, Sweden1Department of Surgery, The Sahlgrenska Academy at Göteborg University, Göteborg, Sweden2Department of Pathology, The Sahlgrenska Academy at Göteborg University, Göteborg, Sweden

(Requests for offprints should be addressed to R Racho El-Akouri, Department of Obstetrics and Gynecology, Sahlgrenska University Hospital,S-413 45 Göteborg, Sweden; Email: [email protected])

Abstract

A method of heterotopic uterine transplantation wasdeveloped in the mouse as a model system for studies ofuterine function and transplant immunology of the uterus.The model involved transplantation of the right uterinehorn and the cervix by vascular anastomosis to a donoranimal with the intact native uterus remaining in situ.F1-hybrids of inbred C57BL/6�CBA/ca (B6 CBAF1)mice of 6–8 weeks of age (n=42) were used. The specificpelvic vascular anatomy of these mice was first examinedby intra-aortal injection of a two-component silicon-rubber curing agent. The surgery of the donor animalinvolved microsurgical isolation of the right uterine hornand the cervix, with preserved vascular supply from theaorta through the right uterine artery. After isolation of theuterine horn with vascular supply and venous drainage,including approximately 3 mm of the inferior vena cavaand aorta, the organ was put on ice. The recipient animalwas prepared by exposing and mobilizing the infrarenalpart of the aorta and the vena cava. The grafted uterus wasplaced in the abdomen on the left side and the aorta andvena cava of the graft were anastomosed end-to-side to theaorta and vena cava of the recipient animal with 11–0

sutures. The total time for these procedures declined withtime and was 125�4 min for the last 28 operations.Viability of the uterus was confirmed, several days later, bydemonstrating a blood flow similar to that of the nativeuterus, and histology of the grafted uterus demonstratednormal morphology, including intact ultrastructure of theendothelial cells. The overall survival rate of the recipientanimals increased with learning from approximately 40%in animals 1–21 to 71% in animals 22–42. The proportionof viable grafts, as judged by normal blood flow andhistology among the surviving mice was 25% in animals1–21 and 87% in animals 22–42. An undisturbed functionof the transplanted uterus horn was finally demonstratedby its ability to implant inserted blastocysts and to carrypregnancy with fetal weight being similar to that of fetusesin the native uterus and controls. In conclusion, this is thefirst report of successful transplantation of the uterus withproven functionality in the mouse. The model should beuseful for many aspects of research in uterine physiologyand pathophysiology.Journal of Endocrinology (2002) 174, 157–166

Introduction

The mouse is considered to be the most important animalwithin many medical research fields, including endocrinol-ogy. The increasing availability of inbred mice strains withtypical features (Wiktor-Jedrzejczak et al. 1990, Pollardet al. 1991, Cohen et al. 1997) as well as of specifictransgenic and knock-out mice (Stewart et al. 1992,Kauma 2000, Zoubina & Smith 2001) has led to a tremen-dous advance in attempts to understand complex physio-logical processes, including reproductive events. The dataaccumulated from experiments in these animal models hasso far been from studies of transgenic or gene-deleted mice,

mostly carrying global upregulations or deletions of specificgenes. It is not known from these animal models whether aspecific uterine phenotype is attributable to secondaryhormonal or other systemic imbalances or is a result of theincreased or deleted expression of the gene in the uterus.Furthermore, there are no uterine tissue-specific pro-moters known today to allow for construction of condi-tional uterine-specific knock-out mice. To be able to studythe effects of a specific mediator in the uterus it would beof great benefit to develop techniques for transplantation ofthe uterus in a mouse model.

During the last decades there have also been tremen-dous developments in the field of organ transplantation.

157

Journal of Endocrinology (2002) 174, 157–1660022–0795/02/0174–157 � 2002 Society for Endocrinology Printed in Great Britain

Online version via http://www.endocrinology.org

The clinical work in this field has mainly been concen-trated on the development of transplantation of centralorgans, such as the heart, the kidney and the liver. Animalmodels have generally been developed prior to introduc-tion of transplantation of specific organs in the human toinvestigate important issues such as tissue rejection and thespecific immunology involved (Lee et al. 1973, Blom &Orloff 1998, Hoerstrup et al. 2000). With the recentdevelopments of more effective immunosuppression tocontrol tissue rejection, the research and clinical fields oftransplantation have also come to involve studies ontransplantation of other organs such as the pancreas and thesmall intestine (Kort et al. 1973, Di Cataldo et al. 1989, Heet al. 1998). Transplantation of reproductive organs hasbeen considered in reproductive medicine for both scien-tific and clinical purposes. Among the female reproductiveorgans, a majority of studies have been conducted to studyvarious aspects of ovarian transplantation, first attempted asa potential treatment for the menopause (Davidson 1912,Nattrass 1915). The animal models used for ovariantransplantation were generally autotransplantation models,with only the ovary replanted (Baird et al. 1976) or theovary replanted together with the oviduct (Winston &Browne 1974) within the same animal. There are fewerreports in the literature concerning transplantation of theuterus. The uterus was transplanted together with theoviduct and ovary as shown in the ewe (Baird et al. 1976),the dog (Yonemoto et al. 1969, Mattingly et al. 1970) andthe macaque monkey (Scott et al. 1971). The techniquefor uterine transplantation was either by direct vascularanastomosis (Yonemoto et al. 1969, Mattingly et al. 1970)or by omentopexy to obtain revascularization (O’Learyet al. 1969, Barzilai et al. 1973). In the more commonlyused small experimental animals, there are early reports onnon-vascular procedures for transplantation of the uterus aspresented in the guinea pig (Bland 1972) and in the rabbit(Confino et al. 1986). Lately, an en-bloc vagino–utero–ovarian vascular anastomosis transplantation technique inthe rat has been presented (Lee et al. 1995). The functionof these transplanted uteri has not been further assessed.

One major obstacle involved in any organ transplan-tation in mice is the small size of the animal. Thus, it isnecessary to use delicate microsurgical techniques forvascular anastomosis. The purpose of the present study wasto describe the surgical technique which we have devel-oped for heterotopic uterine transplantation in the mouseand to evaluate the transplanted organ in terms of tissueblood perfusion and histology. In addition, we presentinitial results demonstrating successful implantation andpregnancy in the transplanted uterus.

Materials and Methods

AnimalsF1-hybrids of inbred female C57BL/6�CBA/ca (B6CBAF1) mice of 6–8 weeks of age (weight 20–25 g) were

purchased from M&B A/S (Ry, Denmark). Animals werehoused with up to ten in each cage and they had freeexcess to water and pelleted food. The temperature in theanimal quarter was between 21 and 23 �C, with a relativehumidity of 50–60%. The animal room was illuminatedbetween 0700 and 1900 h and the mice were adapted tothese conditions for at least 5 days prior to the experiments.The experiments were approved by the local animalethics committee and were carried out according to theprinciples and procedures outlined in the NIH Guide andUse of Laboratory Animals.

Vascular anatomy determination by microdissection andMicrofil injection

Initial experiments were designed to obtain a detailedcharacterization of the abdominal and pelvic vascularanatomy in relation to the uterus on both the arterial andvenous side in the mouse strain used. The purpose of thiswas to achieve the background information necessaryfor further development of the microsurgical techniquesinvolved in the transplantation surgery. To obtain thisinformation, microdissection and Microfil injection (FlowTech, Inc., Carver, MA, USA) were used. Microfil is atwo-component silicon-rubber curing agent that has beenused extensively for visualization of the vasculature ofother sites of the body, such as the rat kidney glomeruli(Norlen et al. 1978) and also in the human for studiesof the vascular supply to the thenar area of the hand(Omokawa et al. 1997).

The mouse was anesthetized by continuous admin-istration of 2% isoflurane (Baxter, Kista, Sweden) in amixture of air (500 ml/min) and O2 (500 ml/min) andplaced in a supine position. A mid-line laparotomy wasthen performed and the aorta and the vena cava wereexposed and mobilized. An operating stereomicroscope(Leica M 651 with a DC 300 digital camera connection;Leica microsystems AB, Sollentuna, Sweden) with mag-nifications from 6�to 40�was used. The inferiormesenteric vessels were cauterized and cut followed byretraction of the colon to the left side of the animal. Theaorta and vena cava were then separated by gentledissection between the vessels at a point just below thebranching of the renal vessels. The ovarian vessels on eachside were ligated by 8–0 suture (nylon monofilament;S&T, Neuhausen, Switzerland). The external iliac vesselswere also ligated at a point 5–6 mm distal to the branchingof the inferior epigastric vessels. The aorta and vena cavawere then ligated just below the branching of the renalvessels. Just inferior to the ligation the aorta and vena cavawere catheterized with polyethylene tubing (diameter0·6 mm; Intramedic; Becton-Dickinson, Parsippany, NJ,USA), fixed by a 8–0 silk thread and connected to a 27gauge cannula. A 5 ml syringe containing 0·154 M NaClsupplemented with heparin sulfate (100 IU/ml; LeoPharma AB, Malmö, Sweden) and xylocaine (0·2 mg/ml;

R RACHO EL-AKOURI and others · Uterine transplantation158

www.endocrinology.orgJournal of Endocrinology (2002) 174, 157–166

Astra Zeneca, Göteborg, Sweden) was gently injectedthrough the aortic tubing. The injection continued untiltransparent fluid flowed from the vena cava.

A two-step procedure was performed to enable visual-ization of both the arteries and veins. The blue Microfilcompound (high viscosity; 350–450 cps) was flushedthrough the cannula on the venous side until all veins in allorgans of the lower part of the abdomen appeared blue.The cannula was thereafter closed with a titanium clip.At that point, red Microfil compound (high viscosity;350–450 centipoise (cps)) was injected through the aorticcatheter. This high viscosity Microfil liquid is too viscousto pass through capillaries and will then just fill up thearterial side of the vascular system. The cannula was closedwith a titanium clip. The mouse was killed by cervicaldislocation immediately after the procedures, and theabdominal and pelvic parts of the animal were cut out andplaced in a plastic bag at 4 �C overnight. To achievetransparency of tissue, the specimen was then placed for24-h periods in 50% glycerine, 75% glycerine and 85%glycerine and finally preserved in 100% glycerine.

Donor operation

The animal was anesthetized by isoflurane (see above),placed on its back and the abdomen was shaved andcleaned with 70% ethanol. A laparotomy was performedvia a mid-line incision from the symphysis pubis to thexiphoid process.

The surgical procedure described below was performedto isolate and transplant the right horn of the uterusincluding the upper part of the cervix. To be able toperform this and to identify all vascular structures anoperating stereomicroscope (see above) was used. Thetransplantation specimen and the vascular anatomy withsites of ligation is presented in Fig. 1. The small intestineswere initially retracted to the left side of the animal andpacked in moistened gauze. The inferior mesenteric artery(Fig. 1) was then cauterized using bipolar diathermy(Coa-Comp Bikoagulator; Instrumenta AB, Billdal,Sweden) at a maximum effect of 2 W. The colon wasdivided just above the rectum and retracted to the side toobtain a free operating field. Both ureters were alsocauterized and cut.

To be able to fully expose the external iliac vessels onthe right side, the inferior epigastric vessels were cauter-ized and cut just above the ilioinguinal ligament. A suture(10–0 nylon monofilament; S&T) was placed under theexternal iliac vessels 3–4 mm distal to the branching of theinferior epigastric vessels and the vessels were ligatedproximally and cauterized distally followed by severance inbetween. A thorough dissection with removal of extravas-cular tissue and cauterization of small vessels surroundingthe external iliac vessels was then performed up to a pointjust distal to the branching of the external pudendal vessels(Fig. 1). The pudendal vessels were cauterized and cut.

The vessels supplying the lower portion of the cervix andthe bladder (superior vesical, inferior vesical, cervical andvaginal vessels) were cauterized to prevent bleeding. Thebladder was then removed.

The right uterine horn was then isolated from the leftuterine horn at a site close to the common uterine cavityby ligations (8–0 suture; S&T). The left uterine horn wasremoved and immediately fixed in Kidney Biopsy FixationSolution (Bie & Berntsen A-S, Rødovre, Denmark) to beused for later histological examination and comparisonwith the transplanted right uterine horn.

The cervix was then transected at the level just belowthe attachment to the bladder. Special caution was takento avoid the vascular branches from the right-sidedhypogastric trunk. The cervix could now be lifted upsomewhat to more easily identify the uterine artery andveins and to attach a titanium clip (Weck Closure SystemsLtd, Resarch Triangle Park, NC, USA) on the superiorgluteal vessels. The left common iliac vessels were thenligated with 10–0 suture on the proximal side and a clip onthe distal side followed by severance in between. Thesevessels were then freed from the underlying tissue up tothe caudal vessels by cauterizing some small dorsal vessels.The lumbar vessels (2–4), which branch dorsally from the

Figure 1 Schematic drawing of vascular anatomy of the abdomenand the pelvis of the mice used. Ligations are indicated by stripedbars. 1=vena cava and aorta; 2=renal vessels; 3=ovarian vessels;4= lumbar vessels; 5=caudal vessels; 6= inferior mesenteric artery;7=common iliac vessels; 8=superior gluteal vessels; 9= inferiorepigastric vessels; 10=external iliac vessels; 11=superficialexternal pudendal vessels; 12= internal iliac (hypogastric) vessels;13=posterior division of internal iliac vessels; 14=umbilicalvessels; 15=anterior division of internal iliac vessels; 16= inferiorvesical vessels; 17=uterine vessels.

Uterine transplantation · R RACHO EL-AKOURI and others 159

www.endocrinology.org Journal of Endocrinology (2002) 174, 157–166

aorta and vena cava, were then cauterized and cut. Theaorta and the vena cava were separated from each other bygentle upward traction on the aorta and blunt dissection inbetween from a point just below the branching of the renalvessels and 3–4 mm in a distal direction. The right uterinehorn was ligated by 8–0 suture at a site between the tip ofthe right uterine horn and the oviduct. On the proximalside of the oviduct, a clip was attached and severance wasperformed in between. The right uterine horn was thendissected free from its attachment to the dorsal peritoneumand folded over to expose the right hypogastric trunk.

The aorta and the vena cava were tied off inferior to theovarian vessels by 8–0 suture to create a complete segmentincluding the inferior aorta and the inferior vena cava withthe vascular connection to the right uterine horn. Smallincisions were then made in the vena cava and in the aortajust below the ligation. A 30 gauge needle connected to a1 ml syringe was then inserted into the incision of the aortato gently flush the specimen with 0·4–0·6 ml ice-cold0·154 M NaCl supplemented with heparin sulfate andxylocain (see above) to expand vessels and prevent throm-bosis. During this procedure it could easily be seenwhether the donor uterus was properly perfused byassessing that the uterus blanched and that clear liquidflowed through the incision of the vena cava. The donoruterus was then placed in and kept in ice-cold 0·154 MNaCl. During this ‘back-table’ preparation the aorta andvena cava stumps were separated, cut straight at their endsand cleaned from extravascular tissue to await the surgicalpreparation of the recipient.

Recipient operation

The animal was anesthetized in the same manner as theorgan donor (see above) and then injected s.c. with 15 IUheparin sulfate. The abdomen was shaved and cleanedwith 70% ethanol and a laparotomy was performedthrough a mid-line incision. The intestines were gentlypacked in a saline-moistened gauze and placed to the left.Loose connective tissue overlying the infrarenal aorta andvena cava were removed to expose these areas of the majorvessels. Two to three lumbar vessels were cauterized andcut to enable further mobilization of the aorta and the venacava. Hemostatic clamps (width 1 mm; S&T) were placedaround both aorta and vena cava at sites just above thebranching of the inferior mesenteric vessels and just belowthe branching of ovarian vessels (Fig. 2). To be able toperform end-to-side anastomosis of the graft aorta to therecipient aorta a longitudinal cut (about 1 mm length) wasmade in the anterior aortic wall to create an opening. Thelumen was flushed with 0·154 M NaCl supplementedwith heparin sulfate and xylocaine. During this procedurefor vascular anastomosis, the graft was moistened andkept at room temperature. The aorta of the graft wasanastomosed to the aorta of the host by an end-to-sideanastomosis using continuous 11–0 nylon microsuture

(S&T) on a 50 µm needle. A total of six to eight stitcheswere used for the aortic connection.

In a similar way, the inferior vena cava of the host wasopened by a longitudinal incision of a maximal length of1·5–2 mm. The lumen of the vena cava was flushed andthe graft was anastomosed in an end-to-side manner withcontinuous 11–0 nylon microsuture.

The sites of the two anastomoses were then wrappedwith a small piece of hemostatic matrix (Surgicel; Ethicon,Brussels, Belgium) to minimize leakage and the twohemostatic clamps were removed. Visible leakage from theanastomoses was minimized by gentle pressure with acotton tip. When hemostasis was satisfactory, the intestineswere repositioned into the abdominal cavity and theabdominal wall was closed in two layers with 6–0 suture.To compensate for fluid loss and for hypoglycemia, 1 ml0·154 M NaCl and 1 ml glucose (50 mg/ml) were thenadministered s.c. The animal was placed under a heat lampuntil fully awake. The mouse received heparin sulfate (5IU) s.c. at 24 h after the operation. General condition andthe laparatomy wound were assessed every day for the firstweek and then twice weekly.

Embryo transfer

In one of the transplanted mice (8 weeks old), the abilityof the transplanted uterus to implant transferred embryos 1week after transplantation was tested. Six blastocysts wereretrieved from a naturally mated B6 CBAF1 female 4 daysafter the vaginal plug was observed. The blastocysts weretransferred to the uterine-transplanted mouse (Hogan et al.1994) 3 days after mating with a vasectomized male. The

Figure 2 Schematic drawing showing sites of hemostatic clampsand heterotopic transplanted graft after end-to-side aorta–aortaand vena cava–vena cava anastomoses.



embryo transfer was made during anesthesia (see above) bya midline laparotomy. Direct entrance through the uterinewall was accomplished with a 30 gauge needle and a glasspipette with the blastocysts was introduced through thehole. Three blastocysts were placed in the native uterusand three blastocysts were placed in the graft. The mousewas killed by cervical dislocation on day 10 after embryotransfer and the number of fetuses was counted.

Laser Doppler flowmetryLaser Doppler flowmetry represents a new method forestimation of tissue perfusion (Oberg 1990). The flow-meter used in this study (Periflux 5000; Perimed, Järfälla,Sweden) consists of a solid-state diode laser with a wave-length of 780 nm and a maximum power output at theprobe tip of 1 mW. The probe used (PF 407) containsoptic fibers for transmission of laser light. The principle oflaser Doppler flowmetry is that a laser light is transmittedto the tissue via a fiber optic probe. When this light hitsmoving blood cells, it undergoes a change in wavelength(Doppler shift). The magnitude and frequency distributionof these changes are directly related to the number andvelocity of blood cells. Measurements presented as arbi-trary perfusion units (PU) are viewed and recorded in aPerisoft software computer program (Perisoft, Järfälla,Sweden).

At 15 and 30 days after the operation the mouse wasanesthetized by isoflurane (see above). The abdomen wasopened through a mid-line incision and an adhesive miniprobe (Perimed) was placed on the surface of the graftuterus and on the native uterus. The probe was also placedon the surfaces of the small intestine, liver and the kidneyfor recordings. The blood flow was recorded for periods of2–3 min and the average blood flow, expressed in PU, wasthen calculated in the Perisoft computer program.

Light and electron microscopyThe recipient mice were killed and the native and donoruteri were quickly removed and fixed in Kidney BiopsyFixation Solution. Tissue pieces were then dehydrated andembedded in paraffin for light microscopy (Nikon lightmicroscope and Coolpix MDC camera) after staining withhematoxylin and eosin. Small pieces of tissue were cut andfixed by immersion in 3% glutaraldehyde in 0·1 M sodiumcacodylate buffer, pH 7·4, postfixed in 2% osmiumtetroxide in the same buffer and, after dehydration inalcohol and propylene oxide, embedded in Agar 100epoxy resin. Thin sections were cut and contrasted withuranyl acetate and lead citrate, on the grids, and examinedin a Philips CM12 electron microscope.

Results

Vascular anatomyDuring the development of the microsurgical technique itwas apparent that the vascular anatomy varied somewhat

from that described in text books. It was also difficult tolocalize all vessels through the microscope. The MicrofilRtechnique enabled us to characterize the pelvic vascularnetwork in detail (Fig. 3), which was of importance for thefurther development of the surgical operative procedure.

Figure 3 Uterine vasculature after Microfil injection on botharterial and venous side demonstrating uterine vascular arcade(a) and details of pelvic vascular anatomy (b). 1=vena cava andaorta; 3=ovarian vessels; 7=common iliac vessels; 9= inferiorepigastric vessels; 10=external iliac vessels; 11=superficialexternal pudendal vessels; 12= internal iliac (hypogastric) vessels;16= inferior vesical vessels; 17=uterine vessels. Scale barrepresents 4 mm.



Modifications of surgery and transplantation

A total of 42 animals was included and out of these 23animals (55%) regained good health. Some animals (n=8)died due to hemorrhage during surgery or during the first48-h post-operative period. The other excluded animals(n=11) were killed since they were found to be paraplegicafter the operation. These paraplegic animals were allamong the first 14 animals operated (Fig. 4). The survivalrate increased from 38% among the first 21 animals to 71%in the last series (animals 22–42). The rate of successfultransplantation, as judged by a well-preserved graft withnormal blood flow, increased from 25% for animals 1–21up to 87% for animals 22–42 (Fig. 4). The time for surgeryincluding donor operation, graft preparation and recipientoperation gradually decreased with experience of theprocedures (Fig. 5). In the animals in which blastocystswere transferred (three blastocysts to native uterus, threeblastocysts to transplanted uterus), three and one normalfetuses were seen in the native and transplanted uterus(Fig. 6) respectively. In the transplanted uterus, oneabsorbed pregnancy was seen.

Morphology

On visual inspection, a normal color, tone and consistencyof transplanted uteri were seen in successful grafts atseveral time-points during the observation period of 8weeks after operation. Light microscopic examination ofthe graft (Fig. 7) after the transplantation demonstrated anormal structure with all layers of the uterine wall and theepithelial layer of the uterine glands exhibiting normalmorphology when compared with donor and native uteri.

Electron microscopy

Grafted tissue as well as controls were examined ultrastruc-turally after 4 weeks. No apparent differences betweenthe groups were observed in the glandular epithelium,muscle wall or blood vessels. Special attention was givento the endothelium in capillary walls (Fig. 8), but noinflammatory injuries or structural changes were observed.

Blood flow

The blood flow recordings demonstrated a similar flowin the graft uterus as compared with the native uterus(Fig. 9). The liver, small intestine and the kidney

Figure 4 Survival after surgery and successful grafts among micethat survived.

Figure 5 Time (means�S.E.M.) for surgery of donor (graftharvesting), preparation of isolated specimen (cold ischemia) andsurgery of recipient (warm ischemia and anastomosis).

Figure 6 Visible fetuses in graft and native uteri. The arrow markssite of resorbing implantation. Scale bar represents 5 mm.



Figure 7 Light microscopic examination of native (a, c, e) vs graft uterus (b, d, e) at 2 weeks (a, b), 4 weeks (c, d) and 8 weeks (e, f) afteroperation: magnification �40. The epithelial cells and the stroma are similar in the graft compared with the native uterus. Scale barrepresents 30 µm.



exhibited higher blood flow as compared with the uterus(Fig. 9).

Discussion

The aim of the present study was to develop a method fortransplantation of the uterus which could be used forfurther studies of various aspects of both uterine physiologyand the immunology of uterus rejection. Although uterinetransplantation has been achieved in other animals such asthe dog (Yonemoto et al. 1969, Mattingly et al. 1970), theewe (Baird et al. 1976) and the rat (Lee et al. 1995) amurine model would have advantages, in large partbecause of the great availability of defined inbred andgenetically engineered mouse strains. Moreover, impor-tant research tools such as recombinant proteins, mono-clonal antibodies and specific assays are extremely welldeveloped for the mouse as compared with otherexperimental animal species.

We chose to develop a transplantation method withvascular anastomosis between the graft and the host. Someprevious reports on uterine transplantation in other speciesused an omental–pedicle graft as a source to obtain uterinerevascularization (O’Leary et al. 1969, Scott et al. 1970). Acomparative study of this omentopexy method and directvascular anastomosis techniques in autotransplantation ofthe uterus of the dog (Barzilai et al. 1973) showed that thenew blood supply obtained through revascularization fromthe omentum may only be sufficient for a short time-period after transplantation, since a large proportion of the

transplanted animals had their uterus replaced by a solidfibrous mass (Barzilai et al. 1973). However, it must beemphasized that the microsurgical technique used in thepresent study is extremely delicate. There is a requirementfor a high magnification operating microscope, since manyof the vessels that are isolated and severed have a diameterwell below 500 µm. The diameters of the aorta and venacava that were anastomosed end-to-side are about 700 µmand 1500 µm respectively in mice of around 25 g bodyweight (Holt et al. 1981).

The surgical procedure which took the longest time wasthe isolation of the graft, where great caution had to betaken to preserve the blood supply to the uterus through-out the operation and also to avoid leakage from the vesselsbranching from the vasculature of the graft. It is evidentfrom the learning curve of the present study that the timesfor this and also the other procedures were substantiallyreduced over time.

The surgery for uterine transplantation seems tobe more complicated than other transplantation pro-cedures used in modern research involving gene-deletedmouse models. Transplantation of ovaries from wild-typeto knock-out mice or vice versa is a widely used tech-nique. This technique is, compared with uterine trans-plantation, rather simple. The removal of ovaries isperformed from the bursal cavity by sharp scissors and thegrafted ovaries can then simply be put into the bursa forsubsequent revascularization with a subsequent satisfactoryreturn to fertility (Yun et al. 1990).

In this first report on uterine transplantation in themouse we decided to develop a heterotopic model, where

Figure 8 Transmission electron microscope details of uterine capillaries in native uterus (a) and graft uterus (b). The endothelial cells aresimilar in the transplanted uterine tissue compared with the native uterus. No signs of rejection or degeneration can be seen. Scale barrepresents 5 µm.



the transplanted uterine horn was grafted alongside therecipient’s uterus. A reason for using this heterotopicmodel was that it would enable us to compare the functionand morphology of the grafted uterus with the nativeuterus within the same animal. We found that bothmorphology and blood flow of the grafted uterus weresimilar to that of the native uterus.

The method of laser Doppler flowmetry to assess graftviability has previously been used for follow-up of smallbowel transplants in mice where it was noticed, innon-rejecting strain combinations, that a similar flow wasseen in graft and native intestine (Dindelegan et al. 2001).The blood flow of the uterus recorded in the present studywas in the same range as that described earlier for theuterus of the rat (Zackrisson et al. 2000) and we also founda similar ratio between uterine blood flow and blood flowof other organs, such as the liver and the kidney.

The transplanted uterus horn showed similar blood flowand morphology to that of the native uterus in spite ofthe fact that this transplanted organ is likely to havedisturbances in innervation and in lymphatic drainage.Moreover, the blood supply to the transplanted uterinehorn is only through one of the two natural sources, theuterine artery and the uterine branch of the ovarian artery.

In the present study, we have demonstrated pregnancyachieved in a uterus transplanted as an isograft. This is anadvance in comparison with previous reports of preg-nancies in the autografted uterus or of the dog (Eraslan

et al. 1966). This is, to our knowledge, the first pregnancyreported in a model not involving autotransplantation(replantation). In the autotransplantation experiments ondogs (Eraslan et al. 1966), successful pregnancy occurred inthree out of eighteen females that were studied.

In these initial experiments we chose a syngeneictransplantation model to avoid rejection. Knowledge aboutthe immunologic response in uterus transplantation inallogeneic models is very limited. Studies with allogenictransplantation of the uterus in rhesus monkeys, using theomental–pedicle technique, showed rejection of the trans-planted uterus during a 2- to 3-week period after thetransplantation (Scott et al. 1971). The histologic appear-ance was that of progressive infiltration of mononuclearimmune cells. Furthermore, transplantation of the uterusin the rhesus monkey revealed that the rejection of theendometrium was more rapid than that of the myo-metrium (Scott et al. 1971). In a recent study involv-ing vagino–utero–ovarian allogeneic transplantation withend-to-side aortic–aortic, vena cava–vena cava vascularanastomosis (Lee et al. 1995) the findings were that ofedema on day 5 and hemorrhagic edema on day 10.

In future attempts to accomplish uterine transplantationin the human, for fertility purposes, in patient groups suchas those with uterine agenesis or patients who haveundergone hysterectomy in early years because of benignor malignant diseases, a thorough study of the rejectionphenomenon is needed. An allogeneic mouse modelwould be the necessary experimental model to accumulateknowledge and it would be feasible to extrapolate theseresults of experiments in the mouse to possible clinicalsituations in the human.

In studies of functions due to altered expression of aspecific gene in an organ, conditional knock-out mousemodels have been used with the expression being coupledto tissue-specific promoters. For example, to achieveselective altered expression in the pancreas or the liver, usecan be made of insulin and albumin promoters respectively(Postic et al. 1999). Concerning the uterus, there is to dateno tissue-specific promoter that can be used and the uterustransplantation model presented in this study may thusoffer novel opportunities for investigators to test thecontribution of a specific gene to the physiology of theorgan. It would, for example, be of great interest to studythe importance of specific gene transcription within theuterus in processes such as decidualization, implantationand parturition.

References

Baird DT, Land RB, Scaramuzzi RJ & Wheeler AG 1976 Functionalassessment of the autotransplanted uterus and ovary in the ewe.Proceedings of the Royal Society: Biological Sciences 192 463–474.

Barzilai A, Paldi E, Gal D & Hampel N 1973 Autotransplantation ofthe uterus and ovaries in dogs. Israel Journal of Medical Sciences 949–52.

Figure 9 Blood flow measured by laser Doppler flowmetry andexpressed as perfusion units (PU). Values are means�S.E.M. ofrecordings from eight to ten animals (SI, L and K) or four animals(N and G). SI=small intestine; L= liver; K=kidney; N=nativeuterus; G=graft uterus: POD=post-operative day.



Bland KP 1972 Implantation and early pregnancy in theautotransplanted uterus of the guinea-pig. Journal of Reproduction andFertility 28 103–104.

Blom D & Orloff MS 1998 A more versatile and reliable method forrenal transplantation in the rat. Microsurgery 18 267–269.

Cohen PE, Zhu L & Pollard JW 1997 Absence of colony stimulatingfactor-1 in osteopetrotic (csfmop/csfmop) mice disrupts estrouscycles and ovulation. Biology of Reproduction 56 110–118.

Confino E, Vermesh M, Thomas W Jr & Gleicher N 1986 Non-vascular transplantation of the rabbit uterus. International Journal ofGynecology and Obstetrics 24 321–325.

Davidson HS 1912 Transplantation of the ovary in the human being:record of three cases. Edinburgh Medical Journal 3 441–449.

Di Cataldo A, Puleo S, Li Destri G, Guastella T, Trombatore G, LaGreca G, Leone F, Latteri F & Rodolico G 1989 Pancreastransplantation in the rat using Lee’s technique: a reliable model inexperimental microsurgery. Journal of Investigative Surgery 2 159–167.

Dindelegan GLH, Nilsson O, Aneman A, Kurlberg G, Lycke N &Olausson M 2001 Mucosal perfusion during acute rejection ofmouse small bowel allograft. Proceedings of the Inaugural Congress ofthe World Society for Reconstructive Microsurgery, pp 567–568. Bologna,Italy: Monduzzi.

Eraslan S, Hamernik RJ & Hardy JD 1966 Replantation of uterus andovaries in dogs, with successful pregnancy. Archives of Surgery 929–12.

He G, Hart J, Thistlethwaite JR Jr & Newell KA 1998 Modifiedsurgical model of paratopic small bowel transplantation in mice.Journal of Surgical Research 80 188–193.

Hoerstrup SP, Stammberger U, Hillinger S, Zund G, Frick R, LachatM & Schmid R 2000 Modified technique for heterotopic rat hearttransplantation under cardioplegic arrest. Journal of InvestigativeSurgery 13 73–77.

Hogan B, Beddington R & Constantini F 1994 Manipulating the MouseEmbryo. New York, USA: Cold Spring Harbor Laboratory Press.

Holt JP, Rhode EA, Holt WW & Kines H 1981 Geometric similarityof aorta, venae cavae, and certain of their branches in mammals.American Journal of Physiology 241 R100–R104.

Kauma SW 2000 Cytokines in implantation. Journal of Reproduction andFertility 55 (Suppl) 31–42.

Kort WJ, Westbroek DL, MacDicken I & Lameijer LD 1973Orthotopic total small bowel transplantation in the rat. EuropeanSurgical Research 5 81–89.

Lee S, Charters AC, Chandler JG & Orloff MJ 1973 A technique fororthotopic liver transplantation in the rat. Transplantation 16664–669.

Lee S, Mao L, Wang Y, D’Silva M, Yoo CH, Wolf P, Chung WS,Takahashi E, Chung DY & Gittes RF 1995 Transplantation ofreproductive organs. Microsurgery 16 191–198.

Mattingly RF, Clark DO, Lutsky II, Huang WY, Stafl A & MaddisonFE 1970 Ovarian function in uteroovarian homotransplantation.American Journal of Obstetrics and Gynecology 108 773–794.

Nattrass JH 1915 Autoplastic ovarian transplantation. Medical Journal ofAustralia 1 49–52.

Norlen BJ, Engberg A, Kallskog O & Wolgast M 1978 Intrarenalhemodynamics in the transplanted rat kidney. Kidney International14 1–9.

Oberg PA 1990 Laser-Doppler flowmetry. Critical Reviews inBiomedical Engineering 18 125–163.

O’Leary JA, Feldman M & Gaensslen DM 1969 Uterine and tubaltransplantation. Fertility and Sterility 20 757–760.

Omokawa S, Ryu J, Tang JB & Han J 1997 Vascular and neuralanatomy of the thenar area of the hand: its surgical applications.Scandinavian Journal of Plastic and Reconstructive Surgery and HandSurgery 99 116–121.

Pollard JW, Hunt JS, Wiktor-Jedrzejczak W & Stanley ER 1991 Apregnancy defect in the osteopetrotic (op/op) mouse demonstratesthe requirement for CSF-1 in female fertility. Developmental Biology148 273–283.

Postic C, Shiota M, Niswender KD, Jetton TL, Chen Y, Moates JM,Shelton KD, Lindner J, Cherrington AD & Magnuson MA 1999Dual roles for glucokinase in glucose homeostasis as determined byliver and pancreatic beta cell-specific gene knock-outs using Crerecombinase. Journal of Biological Chemistry 274 305–315.

Scott JR, Anderson WR, Kling TG & Yannone ME 1970 Uterinetransplantation in dogs. Journal of the Society for GynecologicInvestigation 1 140–148.

Scott JR, Pitkin RM & Yannone ME 1971 Transplantation of theprimate uterus. Surgery, Gynecology and Obstetrics 133 414–418.

Stewart CL, Kaspar P, Brunet LJ, Bhatt H, Gadi I, Kontgen F &Abbondanzo SJ 1992 Blastocyst implantation depends on maternalexpression of leukaemia inhibitory factor. Nature 359 76–79.

Wiktor-Jedrzejczak W, Bartocci A, Ferrante AW Jr, Ahmed-Ansari A,Sell KW, Pollard JW & Stanley ER 1990 Total absence of colony-stimulating factor 1 in the macrophage-deficient osteopetrotic(op/op) mouse. PNAS 87 4828–4832.

Winston RM & Browne JC 1974 Pregnancy following autografttransplantation of Fallopian tube and ovary in the rabbit. Lancet 2494–495.

Yonemoto RH, Du Sold WD & Deliman RM 1969Homotransplantation of uterus and ovaries in dogs. A preliminaryreport. American Journal of Obstetrics and Gynecology 104 1143–1151.

Yun JS, Li YS, Wight DC, Portanova R, Selden RF & Wagner TE1990 The human growth hormone transgene: expression inhemizygous and homozygous mice. Proceedings of the Society forExperimental Biology and Medicine 194 308–313.

Zackrisson U, Mikuni M, Peterson MC, Nilsson B, Janson PO &Brannstrom M 2000 Evidence for the involvement of blood flow-related mechanisms in the ovulatory process of the rat. HumanReproduction 15 264–272.

Zoubina EV & Smith PG 2001 Sympathetic hyperinnervation of theuterus in the estrogen receptor alpha knock-out mouse. Neuroscience103 237–244.

Received 8 March 2002Accepted 16 April 2002



Documents

Heterotopic uterine transplantation by vascular anastomosis in the