京都大学 / 0048 / 新制・課程博士 / 博士(医学) / 甲第18860号 / 医博第3971号 / 新制||医||1008(附属図書館) / 31811 / 京都大学大学院医学研究科医学専攻 / (主査)教授 武藤 学, 教授 平岡 眞寛, 教授 羽賀 博典 / 学位規則第4条第1項該当 / Doctor of Medical Science / Kyoto University / DFAM

Endometriosis is a reproductive disease affecting women in their prime reproductive years characterized by the presence of endometrial glands and stroma in ectopic locations. Animal models have been proven to be indispensable not only for the study of the disease but also to develop better non-invasive diagnostic imaging modalities. A major limitation with the diagnosis of endometriosis is the lack of a specific and sensitive non-invasive diagnostic test. Our objective was to develop a domestic animal model of endometriosis suitable for serial diagnostic imaging procedures and assessment of therapies. Two major studies were conducted to achieve this objective. First study involved in vitro whole tissue-explant culture and surgical induction of endometriosis in dog, pig and sheep to choose the most suitable model. For in vitro co-culture, dog, pig and sheep endometrium was placed on visceral peritoneum for 24 to 72 h to assess the degree of attachment and adhesion characteristics of endometrium (epithelium, glandular and stromal cells). Surgical induction of endometriosis was tested in dog, sheep (n=5 each) and pig (n=4) using autologous endometrial (n=4 grafts per animal) and fat grafts sutured to visceral (urinary bladder surface in dog and pig, uterus in sheep) and parietal (abdominal wall) peritoneum. Sham surgeries were performed in control group animals (dog and sheep n=5, pig n=3) using fat grafts alone. Animals were euthanized between 80-110 days post-surgery. Size, gross characteristics and histopathologic features of endometriotic lesions were recorded. During in vitro explant culture, surface epithelial, stromal and glandular cells of endometrium were capable of attaching to visceral peritoneum within 24 hours with and without an intact layer of mesothelial lining in dog, pig and sheep. The proportion of successful endometrial attachments were greater at 24h compared to 72h (15/18 vs. 7/18, p=0.008; data combined among species) with intermediate attachment at 48h (12/15). Following surgical induction, there was no difference (p>0.05) in proportion of successful tissue grafts placed on serosal surface of visceral vs. parietal peritoneum in dog (10/10 vs. 10/10), pig (7/8 vs. 8/8) or sheep (7/10 vs. 8/10). A variety of outcomes (endometriotic cysts with sero-sangunous fluid, solid lesions, vesicles, absence of lesions) were found. The proportion of cystic lesions was greater (p<0.01) in dog (19/20 grafts) than in pig (8/16) and sheep (5/20). Further, the area of endometriotic lesions at euthanasia was larger (0.89 ± 0.11 cm2) compared to that at the time of surgery (0.50 ± 0.09 cm2) in dog, whereas, the size of lesions decreased by half or more (p<0.05) in pig and sheep. Combined among grafting sites (visceral and parietal peritoneum) and species, a greater proportion (p=0.015) of surgical sites had adhesions in treatment (12/14) versus control group animals (5/13). The wall of majority of endometrial cysts in dogs were characterized by simple cuboidal/columnar epithelium, endometrial glands (normal, dilated and cystic), subepithelial capillary network and presence of stromal and smooth muscle cells. Hemorrhage and/or hemosiderin-laden macrophages were observed in the cystic lesions in dog. Development of a greater proportion of growing lesions in the form of endometriotic cysts in dogs compared to sheep and pig led us to conclude dog as a better suitable domestic animal model for endometriosis than sheep and pig. Second study involved assessing the usefulness and limitations of ultrasonography, magnetic resonance imaging (MRI) and positron emission tomography-computed tomography (PET-CT) in detecting cystic endometriotic lesions in dog and sheep. Surgical induction of endometriois was performed in dogs (n=5) and sheep (n=5) using autologous endometrial grafts (n=4 grafts per animal) and fat grafts sutured to visceral peritoneum (urinary bladder in dogs, uterus in sheep) and parietal peritoneum (ventral abdominal wall). Weekly ultrasonography was performed from Week 1-9 post-surgery and day of euthanasia (Week 14-15). T1 and T2 weighted weighted MRI images (n=2 each for dog and sheep) and PET-CT (n=3 dog, n=1 sheep) using 18F-fluorodeoxyglucose (18F-FDG) as radiolabel was performed between Week 13-15 post-surgery in dog and sheep. Gray-scale B-mode ultrasonography was able to detect endometriotic cysts (0.25-1.75cm) on urinary bladder and abdominal wall in dogs; endometrial grafts Week 1 post-surgery appeared as a homogenous, hypoechoic masses, following which they grew larger with evidence of cyst formation by Week 5. Cysts were undetectable from Week 10-13, whereas they appeared as homogenous masses with a hypoechoic fluid-filled cavity with diffuse hyperechoic echoes and low vascularisation (Color-Doppler imaging) by Week 14-15. In sheep, endometrial grafts were detected as hypoechoic masses Week 1 post-surgery that became smaller until no detectable lesions were visible beyond Week 6-7. Cysts in dogs and sheep appeared hyperintense on T2 and hypointense on T1 weighted images. 18F -FDG PET-CT did not show hypermetabolic activity in endometriotic cysts in dogs and sheep. In conclusion, MRI appeared to provide the most definitive diagnostic images of endometriotic cysts in dogs and sheep, particularly for lesions in sheep which were not evident by ultrasonography. However, ultrasonography was sufficient to characterize most endometriotic cysts in dogs. Further research needs to be carried out to develop specific PET tracers for endometriosis.