Conflict of interest: The authors have declared that no competing interests exist.
Uterine transplantation (UTx) is a potential option for child-bearing in women with uterine infertility. Recovery of uterine function after allogeneic UTx in non-human primates has not been reported. The objective of this study is to establish the functional uterine transplant model in non-human primates.
Uteri of two cynomolgus monkeys were simultaneously removed, cooled at 4°C and perfused with heparin saline. The uteri were interchanged with each other and then orthotopically transplanted. Immunosuppressive protocols included use of three agents (tacrolimus, mycophenolate mofetil and methylprednisolone) in case 1 and two agents (tacrolimus and methylprednisolone) in case 2. Transabdominal ultrasonography, vaginoscopy and biopsy of the transplanted uterine cervix were routinely conducted to monitor rejection after surgery.
The blood concentration of tacrolimus decreased 11 days after surgery and evidence of rejection was found in biopsy of the uterine cervix in both cases. The suspected rejection disappeared 23 days after surgery in case 1 and temporary menstruation resumed at 3 months after surgery. In case 2, blood flow to the uterine artery gradually decreased and the uterus resulted in atrophy due to ischemia, which has been triggered by rejection.
Allogeneic UTx in the cynomolgus monkeys resulted in temporary recovery of menstruation with three immunosuppressants and uterine atrophy with two immunosuppressants. This preliminary experience suggests that recovery of uterine function after allogeneic UTx in non-human primates is possible but more experiments are required.
Assisted reproductive technology (ART) has improved markedly in recent years. However, women with uterine infertility who require hysterectomy due to a malignant uterine tumor, benign disease or post-partum hemorrhaging and those with a congenital defect such as Mayer–Rokitansky–Küster–Hauser syndrome currently have no option of having children, other than adoption and gestational surrogacy. Gestational surrogacy is also restricted due to legal, ethical and religious issues in many countries. Consequently, it is difficult for women with uterine infertility to have children and this may decrease quality of life in young women.
Uterus transplantation (UTx) may allow women with uterine infertility to bear healthy children and have improved quality of life. However, the uterus is not a vital organ, and therefore the procedure remains controversial in humans. The first UTx was conducted in Saudi Arabia in 2000; however, the transplanted uterus developed necrosis and was removed. This led to UTx studies in animal models, combined with recent development of technology for organ transplantation, microvascular anastomosis and immunosuppressant therapy. Basic studies have been conducted in many animals, including non-human primates.
The second UTx in humans was reported in August 2011 in Turkey. After the surgery, periodic menstruation was confirmed with the transplanted uterus, and embryo transfer was attempted from more than 1 year after surgery. Consequently, pregnancy was achieved in April 2013, according to information from the media, although abortion occurred at the first trimester. In September 2012, the group in Sweden conducted two UTx with living donors, as the first procedures between mother and daughter. These data suggest that UTx is now reaching the run-in period to clinical application.
The end-point of UTx differs from reconstruction of other solid organ transplant functions, because the goal is to facilitate pregnancy and delivery of healthy children; however, pregnancy by allogeneic UTx has only been shown in rats and sheep. The next step towards accomplishment of pregnancy and delivery in human UTx is to accumulate data on allogeneic UTx in non-human primates. Several studies of auto-UTx in non-human primates have been conducted and we have reported the first birth in a cynomolgus monkey model after auto-UTx. However, there have been no reports of pregnancy and delivery after allogeneic UTx in primates, and the only performance of allogeneic UTx in non-human primates resulted in assumed failure of resumption of menstruation. Therefore, further accumulation of data on allogeneic UTx in non-human primate models, including pregnancy and delivery, is needed.
This study was performed with the aim of developing a procedure for allogeneic UTx with recovery of uterine function in a cynomolgus monkey primate. We present our preliminary experience of immunosuppressive treatment and rejection in non-human primate models.
This study was conducted in healthy cynomolgus monkeys with regular menstrual cycles. After examining blood types of 23 monkeys, we selected two monkeys with same blood type (case 1, 7 years old, 4.11 kg; case 2, 8 years old, 4.05 kg). Both monkeys had a high degree of polymorphism in the major histocompatibility complex (MHC) gene (Table 1). The study protocol was approved by the Institutional Scientific Evaluation and Review Committee and the Animal Care and Use Committee of the Institute of Primate Research, Shin-Nihon-Kagaku, Kagoshima, Japan (permit no. IACUC720-007, SBL740-008), which is fully accredited by the International Association for Assessment and Accreditation of Laboratory Animal Care (approval no. 001404).
Table 1. Results of Mafa-DNA typing
Animals were anesthetized as previously described.[11, 12]
Two transplantation surgeries using two monkeys, as shown in Figure 1, was planned. We planned to use the uterine artery and ovarian vein (or, if possible, the uterine vein) for arterial and venous vascularization, respectively, in the transplanted uterus. Because the ovary is removed when the ovarian vein is used, only veins of a unilateral ovary were used and the contralateral ovary was retained to maintain ovarian function. The uterus of each monkey was removed almost simultaneously from the abdominal cavity (Fig. 1a).
Back table preparation
Back table preparation was performed as previously described.
After back table preparation, the uteri were interchanged and orthotopically transplanted. In case 1, end-to-end anastomosis of the left uterine artery of the host to the left uterine artery of the uterus of case 2 was carried out by interrupted suture with 12-0 nylon thread (Crownjun). Next, end-to-side anastomosis of the right ovarian vein of the host to the right ovarian vein of the uterus of case 2 was carried out by interrupted suture with 9-0 nylon thread (Crownjun). Clamps for vessels were then released and uterine perfusion started. Subsequently, end-to-end anastomosis of the right uterine artery of the host to the right uterine artery of the uterus of case 2 was carried out by interrupted suture with 12-0 nylon thread. Because the uterine vein was extremely thin, no anastomosis was performed. Thus, in case 1, the uterus was perfused using two arteries and one vein (Fig. 1b).
In case 2, end-to-side anastomosis of the right uterine artery of the host (vascular diameter, 1.2 mm) to the right uterine artery of the uterus of case 1 was carried out by interrupted suture with 11-0 nylon thread (Crownjun). Next, end-to-end anastomosis of the left ovarian vein of the host in the mesosalpinx to the left ovarian vein of the uterus of case 1 was carried out by interrupted suture with 11-0 nylon thread. Clamps for vessels were then released and uterine perfusion started. Subsequently, end-to-end anastomosis of the left uterine artery of the host to the left uterine artery of the uterus of case 1 was carried out by interrupted suture with 11-0 nylon thread, and end-to-end anastomosis of the right uterine vein of the host to the right uterine vein of the uterus of case 1 was carried out by interrupted suture with 11-0 nylon thread. Because the left uterine vein was extremely thin, no anastomosis was performed. Thus, in case 2, the uterus was perfused using two arteries and two veins (Fig. 1b).
To prevent rejection of each transplanted uterus, immunosuppressants were used in the perioperative and postoperative periods. Case 1 was given tacrolimus, a calcineurin inhibitor (Prograf; Astellas Pharma); mycophenolate mofetil (MMF), an antimetabolite (Cellcept; Chugai Pharmaceutical); and methylprednisolone, a corticosteroid (Medrol; Pfizer Japan). Case 2 received tacrolimus (Prograf) and methylprednisolone only.
In case 1, Prograf (0.15 mg/kg) and Cellcept (20 mg/kg) were administrated p.o. prior to surgery. After surgery, Prograf was administrated p.o. at 0.3 mg/kg per day and increased or decreased appropriately based on the blood concentration of the drug and evidence of rejection. Cellcept was administrated p.o. at a dose of 20 mg/kg per day 2 months after surgery and subsequently decreased to 10 mg/kg per day. Medrol was administrated p.o. at 10 mg/day and gradually decreased to 4 mg/day in weekly steps of 2 mg/day.
In case 2, the immunosuppressants other than Cellcept were administrated, similarly to case 1. These immunosuppressants were administrated twice a day at an interval of 12 h. In addition, antibiotic, antiviral, antifungal and antiprotozoal agents as countermeasures against infection and a proton-pump inhibitor to protect against gastric ulcer were administrated p.o. All drugs were administrated using a nasogastric catheter.
The blood concentration of tacrolimus was regularly measured after surgery and the target trough levels were found to be within the planned ranges (postoperatively, 20–30 ng/mL; 2 months after surgery, 15–20 ng/mL; ≥6 months after surgery, 10–15 ng/mL).
To monitor potential rejection after surgery, the size of the transplanted uterus and blood flow in the transplanted uterine artery were determined using transabdominal ultrasonography, and the color and necrosis of the transplanted uterine cervix were observed using vaginoscopy. Biopsy of the transplanted uterine cervix was conducted by clamps (Storz 5Fr; Karl Storz).
The background and surgical details of the two monkeys are shown in Table 2. After vascular anastomosis and release of vascular clamps, the color of both transplanted uteri changed from white to red. Beating of the anastomosed uterine arteries was observed macroscopically in both cases. No uterine congestion was observed and venous return was good in both cases.
Table 2. Background and surgical data for two female cynomolgus macaques that underwent uterus allotransplantation
Time for excision of uterus
5 h 33 min
5 h 56 min
Excised uterine weight
Length of the vagina contained in the excised uterus
Duration of perfusion on the back table
Cooled uterine temperature
Heparinized saline 20 cc
Heparinized saline 20 cc
Uterine total ischemic time
4 h 16 min
3 h 15 min
Cold ischemic time
1 h 13 min
Time required for vascular anastomosis
3 h 41 min
2 h 20 min
Anastomosed blood vessels
2A 1V (bilateral uterine arteries and right ovarian vein)
2A 2V (bilateral uterine arteries and right uterine vein, left ovarian vein)
Changes in blood tacrolimus concentration after surgery are shown in Figure 2.
Assessment of the transplanted uterus using B-Mode or Doppler ultrasonography
The postoperative size of the transplanted uterus in case 1 did not change markedly. In case 2, the size of the transplanted uterus temporarily increased on postoperative day (POD) 23, but subsequently decreased gradually (Table 3). The observation of blood flow in the uterine artery of the transplanted uterus on Doppler echo in case 1 immediately after surgery showed blood flow in the left uterine artery, but not in the right uterine artery. Blood flow in the left uterine artery was good for 3 months after surgery, whereas that in the right uterine artery disappeared. In case 2, blood flow was observed in both uterine arteries immediately after surgery. However, blood flow in the right uterine artery could not be identified and that in the left uterine artery was weak at 1 month after surgery. Then, blood flow could not be observed in both uterine arteries at 2 months after surgery (Table 3).
Table 3. Summary of findings in the transplanted uterus by 3 months after surgery
Blood tacrolimus concentration (ng/mL)
Size of the transplanted uterus (mm) (long axis × anteroposterior diameter)
23.4 × 14.0
24.4 × 17.2
26.8 × 15.0
27.8 × 14.8
26.6 × 10.9
33.3 × 19.1
17.5 × 7.9
14.8 × 5.3
Observation of blood flow in the uterine artery
Color of cervix with vaginoscopic observation
Light yellow, light pink
Histopathological findings by biopsy
Inflammatory cell infiltration in the interstitium and epithelia
Apoptosis of basal keratinocytes
Capillary vessels with foam cells
Little inflammatory cell infiltration in epithelia
Reactive changes in epithelia
Inflammation surrounding vessels in the interstitium
Swollen endothelia without endotheliitis
Slight inflammatory cell infiltration
Normal vascular changes
Glycogen-rich stratified squamous epithelia
No inflammatory cell infiltration
Normal vascular changes
Glycogen-rich stratified squamous epithelia
Keratinized materials with many bacteria colonies and neutrophils
Desquamated epithelial small fragments
Reactive changes in epithelia
Moderate inflammation in the interstitium
Fibrous changes with hyalinization and hemosiderosis
No epithelial lining
No reactive changes
Removal of Uterus
Vaginoscopic observation of the uterine cervix and histological evaluation by biopsy
The transplanted uterine cervix had a pink color when observed transvaginally immediately after surgery. A biopsy was conducted as a control (Fig. 3A). On POD 11, on which the blood tacrolimus concentration had decreased, the color of the uterine cervix was black and rejection was suspected in both cases (Fig. 3B). In case 1, a biopsy gave the histopathological findings shown in Figure 3(C) and Table 3. Immunohistochemical findings showed that CD8-positive lymphocytes were mainly present in lymphocytic infiltration in the epithelium and interstitium, and that the number of CD20-positive lymphocytes was small (Fig. 4a). In case 2, in contrast, the findings were small fragments in stratified squamous epithelia and keratinized material with many bacterial colonies and neutrophils; therefore, cervical interstitium could not be sampled (Fig. 3C). CD8-positive lymphocytes were also observed in delaminated epithelium, but no CD20-positive lymphocytes were found. These histopathological and immunohistochemical findings in both cases were consistent with an acute rejection response.
Complication of bacterial infection in the uterine cervix was suspected in both cases and transvaginal lavage and administration of an antibiotic agent were implemented. On POD 23, on which the tacrolimus concentration was high, case 1 showed an improved uterine cervix with a pink color, but in case 2 uterine stump diastasis and a light yellow vaginal secretion indicated suspected continued infection. In case 1, pathological findings confirmed that thick keratinized materials and bacteria had disappeared and slight inflammatory cell infiltration was found in epithelia. Reactive changes were found in the stratified squamous epithelia, together with inflammation of lymphocytes and neutrophils surrounding vessels in the interstitium. Swollen endothelial cells were observed, but there were no findings of endotheliitis (Fig. 5a,b). Immunohistochemical findings showed only mild infiltration of CD8-positive lymphocytes in the epithelium. The interstitium showed similar amounts of CD20-positive and CD8-positive lymphocytes, showing non-specific inflammation (Fig. 4b). These results indicate that rejection had resolved and only chronic inflammation remained. In case 2, stratified squamous epithelia were almost eliminated and severe erosion and moderate inflammation in the interstitium were observed, mainly with the presence of lymphocytes and neutrophils (Fig. 5c). In lymphocytes of the interstitium, the level of CD8-positive cells was slightly higher than that of CD20-positive cells, showing possible effects of rejection. On POD 67, by which time the blood tacrolimus concentration had stabilized, the transplanted uterine cervix had a good pink color in case 1, but was white in case 2. In case 1, pathological findings showed slight inflammatory cell infiltration in the epithelium or interstitium, and vascular changes were normal. Glycogen-rich stratified squamous epithelia were found, suggesting an improved estrogen level (Fig. 5d,e). In case 2, fibrous tissues with hyalinization and hemosiderosis alone were found and no epithelial lining or reactive changes were observed (Fig. 5f, Table 3). In both cases, there was no significant infiltration of lymphocytes found histologically that suggested rejection.
Overall outcome and cyclicity in case 1
In case 1, menstruation resumed at 3 months after surgery. However, this was temporary and amenorrhea was subsequently observed. No response occurred in the uterus after administration of estrogen and progesterone, but no evidence of rejection was found in biopsy tissues of the cervical region. In echo findings obtained 6 months after surgery, the size of the uterus had not changed, but blood flow in the left uterine artery could not be detected. Thus, surgery was performed 7 months after the first surgery to remove the uterus. The uterus was highly adhesive to the bladder and abdominal wall, and similar conditions were observed around the right adnexa (Fig. 6a). Although the size of the uterus was normal, the surface was whitish (Fig. 6b). It was difficult to perform separate identification of the uterine artery due to adhesion. No visual or histopathological abnormalities were found in the removed right ovary and transplanted oviduct (Fig. 6c). In histopathological findings of the uterus, there was no endometrial tissue in the intrauterine cavity and the interstitium in almost all layers of the uterine wall showed hyaline degeneration, excluding the part close to the serous membrane, (Fig. 6d). No histopathological findings suggested a rejection response in the uterus, including in the transplanted oviduct.
Overall outcome and cyclicity in case 2
In case 2, menstruation did not resume and atrophy was found in ultrasonography at 3 months after surgery. Therefore, the uterus was removed after laparotomy. Severe adhesion was found in the pelvis and the uterus was adhered with the rectum and the bladder, with atrophy in the funicular region. Severe adhesion was also found in the region crossing the ureter and uterine artery. Beating of the uterine artery was observed on the pelvic side of the adherent site, but not on the uterine side. Uterine stump diastasis was observed with complication of infection (Fig. 7a). Pathological findings of the resected uterus showed uterine atrophy, no epithelium (endometrium), and fibrosis with hemosiderosis and calcification (Fig. 7b). Immunostaining showed a non-specific inflammatory response with slight infiltration of CD8-positive and CD20-positive lymphocytes in the interstitium, and no rejection response. No marked thrombus was found in the uterine artery. The left ovary that was left in the pelvic cavity had follicles and corpora lutea and was normal.
In this study, we conducted allogeneic UTx in cynomolgus monkeys. Allogeneic UTx in non-human primates has only been reported to date, although similar procedures have been performed in several animals. The current study was limited to only two cynomolgus monkeys, but we believe that a detailed evaluation and discussion, including MHC typing and immunohistochemical analysis, is required for each primate in these experiments, similarly to that performed for humans. Primates are physiologically and anatomically similar to humans, and thus our results are potentially important for clinical application of UTx in humans.
Postoperative management for primates differs from that for humans. Because the appropriate concentration of tacrolimus in organ transplantation in cynomolgus monkeys is generally higher than that in humans, we used a higher concentration than that used in humans. It is also difficult to perform continuous infusion, which made it more difficult to control the blood tacrolimus concentration, which had to be stabilized by p.o. administration. Blood tacrolimus decreased 1–2 weeks after surgery due to anorexia, and gastrointestinal absorption was also poor after surgery, with evidence of possible rejection found in both cases. Because low blood concentrations and rejection were observed, the dose of immunosuppressants was increased. The general condition and appetite then gradually improved and at 3 weeks the tacrolimus level rapidly increased, perhaps due to enhanced gastrointestinal absorption of the drug. Thus, it was extremely difficult to control the blood concentrations of oral tacrolimus in the cynomolgus monkeys. Furthermore, a limitation of the study was that tacrolimus could not be determined in the test facility and this test was commissioned to an external institution. Consequently, the results had a time lag of several days. This caused further difficulty with the blood concentration control.
Rejection diagnosis in solid organ transplantation is mostly performed by biopsy. However, there is no clear procedure for monitoring rejection in UTx. In transplantation of other organs, information on organ dysfunction is obtained from blood samples. However, the uterus is not a vital organ and blood tests cannot be used to determine rejection. Therefore, we used Duplex/Doppler echo and pathological findings from biopsy of the uterine cervix to monitor possible rejection. Echo findings show whether blood flow in the uterine artery after microvascular anastomosis is decreased by stenosis or thrombus. In case 2, echo immediately after surgery showed blood flow in the right and left uterine arteries, but flow in the right uterine artery could not be detected after 1 month and there was no flow in both uterine arteries after 2 months, because case 2 did not recover from rejection. Moreover, temporal enlargement of the uterus was observed in case 2 on POD 23. This may be a mechanism of rejection similar to that of renal enlargement observed in renal transplantation.
Pathological findings show that both animals had initial rejection. However, case 1 returned to normal conditions after rejection resolved, whereas case 2 had uterine atrophy due to ischemia, which might have been triggered by rejection. In both cases, the concentration of tacrolimus decreased, causing acute rejection, but in case 1 acute rejection was improved by administration of MMF, while in case 2 the lack of administration of MMF resulted in significant reactions that caused ischemia of the uterus and epithelial detachment, and the effects of acute rejection were not avoided. Therefore, the lack of administration of MMF might have been a cause of the failure to overcome acute rejection, and thus administration of three immunosuppressants, including MMF, may be a favorable protocol for maintenance therapy in future UTx experiments in primate models.
In case 1, uterine nutrition was given mainly from the left uterine artery and right ovarian vein, and these vessels and three immunosuppressants facilitated recovery of menstruation. However, menstruation did not continue despite no subsequent observation of a rejection response. This may be due to insufficient blood flow from the uterine artery to the uterus due to severe adhesion of a region surrounding the uterus.
Because heparinized saline was used as perfusate and the ischemic time was 3 h or longer, ischemia–reperfusion injury might have been one of the causes of the failure of recovery of uterine function. However, we also used heparinized saline for cynomolgus monkeys with an ischemic time of 4 h in an examination of autologous transplantation of the uterus, with the result of successful pregnancy and childbirth. Thus, we consider that ischemia–reperfusion injury was not a major cause of the failed recovery of uterine function. However, a protective preservation solution may minimize problems caused by ischemic reperfusion and further studies of the perfusion solution are required. Studies in humans have shown that uterine myometrial tissue can endure cold ischemia for 6–24 h if stored in protective preservation solution, based on histological findings.[13-15]
One advantage of use of cynomolgus monkey as a primate transplantation model is that the monkey is physiologically and anatomically similar to humans. Therefore, the results should be meaningful for clinical applications in humans. However, there are also several disadvantages. The body size is the same as human infants and this lengthens the surgery time, the animal cost is significant, and postoperative echo and biopsy require sedation with anesthesia. Also, because the pelvis is highly adhesive after surgery, spontaneous pregnancy is not expected due to adhesive tubal obstruction; therefore, ART is required for pregnancy. Embryo transfer is carried out transvaginally in the uterus in humans, whereas the uterine cervix of the cynomolgus monkey is extremely bent, which makes transvaginal embryo transfer technically difficult.
In summary, allogeneic UTx in two cynomolgus monkeys resulted in temporary recovery of menstruation in one animal treated with three immunosuppressants and uterine atrophy due to continuous ischemia caused by rejection in the other animal after treatment with two immunosuppressants. In contrast to transplantation of other organs for recovery of organ function, the ultimate objective of UTx is pregnancy and delivery of healthy children. Thus, in this study, the preliminary goal was recovery of uterine function. The surgical procedure for UTx, immunosuppression, diagnosis of rejection, ischemic reperfusion injury, changes in the immune mechanism during pregnancy and evaluation of uterine blood flow all require further optimization. Further accumulation of data from animal models, including pregnancy and delivery, is needed to establish clinical application of UTx in humans, although UTx in humans has become a clinical reality. Therefore, the preliminary experience in non-human primates reported here is an important step towards further UTx basic research and clinical application of UTx in humans.
We are grateful to Dr Timothy Shim, Dr Kazuki Kikuchi and Dr Kensuke Tashiro (Department of Plastic and Reconstructive Surgery, Graduate School of Medicine, University of Tokyo) for help with surgery; to Hirohito Kato, Nobuyoshi Yamashita, Yoshiro Nishida, Kotaro Hanaki, Ryuichi Katagiri, Tomoko Shimonosono and Syuzo Koyama (Shin Nippon Biomedical Laboratories) for experimental support; to Noriko Kagawa (the chief of Repro Self Bank, Japan) for her advice with hormonal examination; to Tomoharu Mine and Yuhei Shigeta (IMI) for technical assistance and to Hiroshi Suzuki (Department of Pathology, School of Medicine, Keio University) for technical assistance with the immunohistochemical analysis. This study was supported by the Strategic Research Foundation Grant-aided Project for Private Universities from Ministry of Education, Culture, Sport, Science, and Technology, Japan (MEXT), a Keio University Grant-in-Aid for Encouragement of Young Medical Scientists, Kanzawa Medical Research Foundation, Akaeda Medical Research Foundation, Inamori Research Foundation and the Program for the Next Generation of World-leading Research of the Japanese Cabinet Office (LS039). The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.