IntroductionEndometriosis is an estrogen-dependent chronic inflammatory condition mostly affecting women of reproductive age and is associated with chronic pelvic pain and infertility. Even after long three hundred years, most of the literatures [1Aromatase and other steroidogenic genes in endometriosis: translational aspects., 2Khan K.N. Kitajima M. Hiraki H. Fujishita A. Sekine I. Ishimaru T. et al.Immunopathogenesis of pelvic endometriosis: role of hepatocyte growth factor, macrophages and ovarian steroids., 3Pathogenesis and pathophysiology of endometriosis.] still claim that pathogenesis and/or pathophysiology of endometriosis is unclear. We came to learn so far some established hypothesis and regulatory factors supporting the development and maintenance of endometriosis [1Aromatase and other steroidogenic genes in endometriosis: translational aspects., 3Pathogenesis and pathophysiology of endometriosis.]. Recently it has been demonstrated that besides hormonal regulation, both initial and secondary inflammatory mediators are known to be involved in the growth of endometriosis [2Khan K.N. Kitajima M. Hiraki H. Fujishita A. Sekine I. Ishimaru T. et al.Immunopathogenesis of pelvic endometriosis: role of hepatocyte growth factor, macrophages and ovarian steroids., 4Khan K.N. Kitajima M. Hiraki H. Fujishita A. Sekine I. Ishimaru T. et al.Toll-like receptors in innate immunity: role of bacterial endotoxin and toll-like receptor 4 (TLR4) in endometrium and endometriosis., 5Khan K.N. Kitajima M. Hiraki K. Yamaguchi N. Katamine S. Matsuyama T. et al.Escherichia coli contamination of menstrual blood and effect of bacterial endotoxin on endometriosis., 6Khan K.N. Fujishita A. Hiraki K. Kitajima M. Nakashima M. Fushiki S. et al.Bacterial contamination hypothesis: a new concept in endometriosis.]. However, it is difficult to uniformly explain the pathogenesis of endometriosis by a single factor. Therefore, even after many years with enormous publications, most of the literatures tag endometriosis as an enigmatic disease. The recurrence of pain and lesions is still occurring after emergence of effective medical and surgical therapies.In an attempt to search additional factor, recently we established “bacterial contamination hypothesis”, a new concept in endometriosis [5Khan K.N. Kitajima M. Hiraki K. Yamaguchi N. Katamine S. Matsuyama T. et al.Escherichia coli contamination of menstrual blood and effect of bacterial endotoxin on endometriosis., 6Khan K.N. Fujishita A. Hiraki K. Kitajima M. Nakashima M. Fushiki S. et al.Bacterial contamination hypothesis: a new concept in endometriosis.]. As an initial inflammatory mediator, bacterial endotoxin or lipopolysaccharide (LPS) has been found to regulate Toll-like receptor 4 (TLR4)-mediated growth of endometriosis [5Khan K.N. Kitajima M. Hiraki K. Yamaguchi N. Katamine S. Matsuyama T. et al.Escherichia coli contamination of menstrual blood and effect of bacterial endotoxin on endometriosis., 6Khan K.N. Fujishita A. Hiraki K. Kitajima M. Nakashima M. Fushiki S. et al.Bacterial contamination hypothesis: a new concept in endometriosis.]. In order to identify the source of LPS, we demonstrated that menstrual blood of women with endometriosis was highly contaminated with Escherichia coli (E.coli) with a consequent elevation in the concentrations of endotoxin (LPS) in menstrual and peritoneal fluids [[5]Khan K.N. Kitajima M. Hiraki K. Yamaguchi N. Katamine S. Matsuyama T. et al.Escherichia coli contamination of menstrual blood and effect of bacterial endotoxin on endometriosis.]. We found higher concentrations of prostaglandin E2 (PGE2) in menstrual and peritoneal fluid of women with endometriosis [[7]Khan K.N. Kitajima M. Yamaguchi N. Fujishita A. Nakashima M. Ishimaru T. et al.Role of prostaglandin E2 in bacterial growth in women with endometriosis.]. Using in vitro culture system we showed that PGE2 was able to significantly increase the growth of E.coli and administration of levofloxacin, a broad-spectrum antibiotic, significantly suppressed PGE2-promoted growth of E.coli [7Khan K.N. Kitajima M. Yamaguchi N. Fujishita A. Nakashima M. Ishimaru T. et al.Role of prostaglandin E2 in bacterial growth in women with endometriosis., 8Khan K.N. Fujishita A. Kitajima M. Hiraki K. Nakashima M. Masuzaki H. Intra-uterine microbial colonization and occurrence of endometritis in women with endometriosis.]. Our subsequent study further demonstrated increased intrauterine microbial colonization (IUMC) using endometrial samples in bacteria culture system and concurrent occurrence of chronic endometritis in women with endometriosis [[8]Khan K.N. Fujishita A. Kitajima M. Hiraki K. Nakashima M. Masuzaki H. Intra-uterine microbial colonization and occurrence of endometritis in women with endometriosis.]. With the assumption that bacteria culture system has some limitation, because only a fraction of the bacteria present in most microbial ecosystems are amenable to propagation in the laboratory [[9]Hugenholtz P. Goebel B.M. Pace N.R. Impact of culture-independent studies on the emerging phylogenetic view of bacterial diversity.]. Bacteria in complex microbial communities can be identified by characterizing their ribosomal RNA genes (rDNA), a molecular approach that has the advantage of detecting fastidious or cultivation-resistant organisms [[10]A molecular view of microbial diversity and the biosphere.]. Therefore, more recently we re-confirmed our findings of bacteria culture system using Illumina Miseq system and found increased microbial colonization in intrauterine environment and cystic fluid of women with and without ovarian endometriosis [[11]Khan K.N. Fujishita A. Masumoto H. Muto H. Kitajima M. Masuzaki H. et al.Molecular detection of intrauterine microbial colonization in women with endometriosis.]. While bacteria culture result was negative, 16S rDNA metagenome assay could detect significantly higher percentage of Streptococcaceae and Staplylococcaceae in the cystic fluid collected from women with ovarian endometrioma comparing to that in cystic fluid collected from non-endometrioma cysts [[11]Khan K.N. Fujishita A. Masumoto H. Muto H. Kitajima M. Masuzaki H. et al.Molecular detection of intrauterine microbial colonization in women with endometriosis.].Collectively, our findings indicated the occurrence of subclinical infection (without overt symptoms) in intrauterine environment and in the cystic fluid of ovarian endometriosis. If this silent uterine infection persists for long time, it may further aggravate the progression of endometriosis and associated symptoms. Additionally, the fertility outcomes of patients with chronic endometritis significantly improved after antibiotic treatment, supporting the influence of microorganisms in endometriosis [12Cicinelli E. Matteo M. Tinelli R. Pinto V. Marinaccio M. Indraccolo U. et al.Chronic endometritis due to common bacteria is prevalent in women with recurrent miscarriage as confirmed by improved pregnancy outcome after antibiotic treatment., 13Garcia-Penarrubia P. Ruiz-Alcaraz A.J. Martinez-Esparza M. Marin P. Machado-Linde F. Hypothetical roadmap towards endometriosis: prenatal endocrine-disrupting chemical pollutant exposure, anogenital distance, gut-genital microbiota and subclinical infections.]. With these findings in mind, we speculated one unaddressed question, could it be useful in clinics if we add one broad-spectrum antibiotic either alone or in combination with gonadotropin releasing hormone agonist (GnRHa) treatment in order to decrease subclinical intrauterine infection? This could be clinically useful to reduce intrauterine microbial colonization and resulting occurrence of chronic endometritis. Addition of an antimicrobial agent may also reduce the risk of subsequent occurrence or recurrence of endometriosis once GnRHa treatment is stopped thereby may improve reproductive outcome in women with endometriosis. In clinical practice, GnRHa is commonly used for the alleviation of pain, decreasing the risk of intraoperative bleeding and/or for reducing the size of ovarian endometrioma and uterine myoma [[14]Khan K.N. Fujishita A. Koshiba A. Ogawa K. Mori T. Ogi H. et al.Expression profiles of E/P receptors and fibrosis in GnRHa-treated and -untreated women with different uterine leiomyomas.].This hypothesis is supported by one recent study using animal model of endometriosis. In mice treated with broad-spectrum antibiotic, ectopic lesions were significantly smaller with reduction of inflammatory response in pelvis than in those from vehicle-treated mice [[15]Chadchan S.B. Cheng M. Parnell L.A. Yin Y. Schriefer A. Mysorekar I.U. et al.Antibiotic therapy with metronidazole reduced endometriosis disease progression in mice: a potential role for gut microbiota.]. The authors indicated that antibiotic therapy reduces endometriosis progression in mice, possibly by reducing specific gut bacteria [[15]Chadchan S.B. Cheng M. Parnell L.A. Yin Y. Schriefer A. Mysorekar I.U. et al.Antibiotic therapy with metronidazole reduced endometriosis disease progression in mice: a potential role for gut microbiota.]. However, these findings were not translated in human endometriosis. The clinical information on the effect of antibiotic with or without GnRHa on IUMC and its consequent effect on inflammation, cell proliferation and angiogenesis in biopsy samples collected from women with and without endometriosis is unknown. These biological findings could be clinically important in order to aid in the development of personalized treatment strategy.

Therefore, we investigated the following issues using endometrial and endometriotic samples derived from untreated- and levofloxacin (LVFX)-, GnRHa-, LVFX + GnRHa-treated women with and without endometriosis: (a) detection of bacteria in endometrial samples with the use of molecular method, (b) occurrence rate of chronic endometritis in endometria derived from treated and untreated women, (c) changes, if any, in the pattern of tissue inflammation, cell proliferation, and angiogenesis in endometrial and endometriotic biopsy samples collected from treated and untreated women, (d) any histological change of ovarian endometrioma.

DiscussionIn this prospective non-randomized observational study, we demonstrated for the first time that treatment with a single dose of broad-spectrum antibiotic (levofloxacin, LVFX) with or without GnRHa was able to significantly decrease a proportion of bacterial genera in endometrial samples collected from women with and without endometriosis. These findings were coincided with decreased tissue inflammation, cell proliferation and angiogenesis in endometria and endometriotic lesions with remarkable histological improvement of ovarian endometrioma. These findings were based on our serial study for the last 10 years that bacterial contamination indeed occurs in menstrual blood/endometrial samples collected from women with endometriosis associated with increased occurrence of chronic endometritis [5Khan K.N. Kitajima M. Hiraki K. Yamaguchi N. Katamine S. Matsuyama T. et al.Escherichia coli contamination of menstrual blood and effect of bacterial endotoxin on endometriosis., 8Khan K.N. Fujishita A. Kitajima M. Hiraki K. Nakashima M. Masuzaki H. Intra-uterine microbial colonization and occurrence of endometritis in women with endometriosis., 11Khan K.N. Fujishita A. Masumoto H. Muto H. Kitajima M. Masuzaki H. et al.Molecular detection of intrauterine microbial colonization in women with endometriosis.]. Using bacteria culture method and 16S rDNA metagenome assay, we previously demonstrated significantly increased intrauterine bacterial colonization in women with endometriosis comparing to control women. While we previously analyzed bacterial family [[11]Khan K.N. Fujishita A. Masumoto H. Muto H. Kitajima M. Masuzaki H. et al.Molecular detection of intrauterine microbial colonization in women with endometriosis.], here we analyzed bacterial genera by 16S rDNA metagenome assay in both control women and women with endometriosis. In addition to individual effect of LVFX on some target bacteria, addition of LVFX with GnRHa was able to significantly decrease the prevalence rate of Acidibactor and Bradyrhizobium in control women and Acidibactor, Atopobium, Bradyrhizobium in women with endometriosis. The differential response of LVFX and GnRHa to suppress a proportion of analyzed microbiota, and not all, may be due to the difference in their efficacy. Another new finding in this study is that occurrence rate of chronic endometritis was significantly decreased after combined treatment with GnRHa and LVFX.These findings are clinically important, because decrease in IUMC with consequent reduction in the risk of chronic endometritis may improve the adverse reproductive outcome in women suffering from endometriosis. Current treatment strategies, including hormonal therapy and surgery, have significant side effects and do not prevent recurrence. Treatment with an antimicrobial agent either alone or in combination with GnRHa in human endometriosis could be a promising therapeutic approach in ameliorating the occurrence or recurrence of endometriosis. Our findings in human study are supported by a recent elegant study in mice [[15]Chadchan S.B. Cheng M. Parnell L.A. Yin Y. Schriefer A. Mysorekar I.U. et al.Antibiotic therapy with metronidazole reduced endometriosis disease progression in mice: a potential role for gut microbiota.]. In this animal study, the authors found that in mice treated with broad-spectrum antibiotics particularly with metronidazole, endometriotic lesions were significantly smaller with fewer proliferating cells and reduction of inflammatory response than those in vehicle-treated mice [[15]Chadchan S.B. Cheng M. Parnell L.A. Yin Y. Schriefer A. Mysorekar I.U. et al.Antibiotic therapy with metronidazole reduced endometriosis disease progression in mice: a potential role for gut microbiota.]. We have clinically replicated these findings in human endometriosis by our current findings with both antibiotic and hormonal treatment.The 16S rDNA metagenome analysis revealed the existence of various microbiota in endometrial samples of women with and without endometriosis and an individual variability was observed in the distribution of microbiota regardless of treatment. Alpha-diversity analysis confirmed that microbial colonization in uterine environment differs between control women and women with endometriosis with higher prevalence rate in endometriosis and their decrease after individual treatment. Intrauterine microbial colonization of some bacterial genera in women with endometriosis was significantly higher than in control women. This might be due to impaired innate immune system of the female reproductive tract or by overcoming the antimicrobial host defense capacity of Lactobacilli, thereby allowing these major pathogens to persist in intrauterine environment [28Wira C.R. Fahey J.V. Sentman C.L. Pioli P.A. Shen L.i. Innate and adaptive immunity in female genital tract: cellular responses and interactions., 29

C.R. Wira J.V. Fahey M. Ghosh M.V. Patel D.K. Hickey D.O. Ochiel Sex hormone regulation of innate immunity in the female reproductive tract: the role of epithelial cells in balancing reproductive potential with protection against sexually transmitted pathogens 63 6 2010 544 565

]. Although did not reach significant, the population of Lactobacilli was relatively higher in untreated women and after LVFX or GnRHa treatment in control women and women with endometriosis. The critical question remains to be addressed how bacterial population in endometrial samples and its subsequent entry into pelvis may be involved in the pathophysiology of endometriosis. Different ligands from Gram+ bacteria (peptidoglycan/lipopeptide) and Gram- bacteria (LPS) upon their entry in pelvic cavity during retrograde menstruation may trigger TLR2/TLR6 and TLR4 signaling, respectively, in inducing pro-inflammatory response in pelvis and growth of endometriosis. These phenomena of bacterial ligands/TLRs engagement of innate immune system have already been described by previous studies from our laboratory (4-6, 8, 30). Levofloxacin treatment with or without GnRHa may either suppress or reverse this cascade involved in the growth or progression of endometriosis. In fact, we found that LVFX with or without GnRHa treatment significantly decreased tissue infiltration of CD68+ macrophages, Ki-67-stained gland cells/stromal cells and CD31-stained micro-vessel density in the endometria derived from women with and without endometriosis.Local inflammation is one of the major factors that may be involved in the progressive growth of endometriosis and generation of variable degrees of pain manifestations. It has been reported that macrophages drive lesion growth and angiogenesis in mice [31Bacci M. Capobianco A. Monno A. Cottone L. Di Puppo F. Camisa B. et al.Macrophages are alternatively activated in patients with endometriosis and required for growth and vacularization of lesions in a mouse model of disease., 32Lin Y.J. Lai M.D. Lei H.Y. Wing L.Y. Neutrophils and macrophages promote angiogenesis in the early stage of endometriosis in a mouse model.] and enhance IL-1β signaling in response to inflammasome activation, which also promotes endometriotic angiogenesis [[33]Lebovic D.I. Bentzien F. Chao V.A. Garrett E.N. Meng Y.G. Taylor R.N. Induction of an angiogenic phenotype in endometriotic stromal cell cultures by interleukin-1beta.]. Arresting this inflammatory pathway by LVFX either alone or in combination with GnRHa may be important to decrease the growth/progression of endometriosis and to alleviate pain manifestations in women suffering from this disease. In fact, we found that in addition to decreasing CD68-, Ki-67-, and CD31-stained cells in endometria, LVFX treatment with or without GnRHa also significantly decreased the expression of COX2, a rate limiting enzyme for prostaglandin production, and fibronectin, one of the potential cell-cell adhesion molecules, in the endometria of women with and without endometriosis. We previously demonstrated that laminin and fibronectin, two cell adhesion molecules, are highly expressed in endometrial gland cells and stromal cells that were collected during the menstrual phase [[6]Khan K.N. Fujishita A. Hiraki K. Kitajima M. Nakashima M. Fushiki S. et al.Bacterial contamination hypothesis: a new concept in endometriosis.]. The expressions of these adhesion molecules were closely associated with higher expression of integrin-α3 and integrin-α6, the respective receptors of these adhesion molecules in peritoneal mesothelial cells that were also collected during the menstrual phase of similar women with endometriosis [[6]Khan K.N. Fujishita A. Hiraki K. Kitajima M. Nakashima M. Fushiki S. et al.Bacterial contamination hypothesis: a new concept in endometriosis.]. These findings are reasonable to speculate that if we can decrease expression of fibronectin by treatment with either LVFX or GnRHa, then we can decrease the risk of disease development in women with early endometriosis and their time-dependent progression. On the other hand, decreased expression of COX2 in response to LVFX or GnRHa may be related to the decreased production of pain mediators such as prostaglandins and consequent alleviation of pain symptoms in women with endometriosis. Therefore, our current findings of decreased intrauterine microbial colonization in response to LVFX and/or GnRHa have some biological and clinical significance for future application.Mechanistic basis of decreased bacteria profiles in response to LVFX and GnRHa in our current study is unclear. It could be due to decreased GnRHa-mediated local chemokine levels to recruit bacteria in uterine environment or direct bacteriostatic and/or bacteriolytic effect of LVFX [34Piscitelli S.C. Spooner K. Baird B. Chow A.T. Fowler C.L. Williams R.R. et al.Pharmacokinetics and safety of high-dose and extended-interval regimens of levofloxacin in human immunodeficiency virus-infected patients., 35Valore E.V. Park C.H. Quayle A.J. Wiles K.R. McCray P.B. Ganz T. Human beta-defensin-1: an antimicrobial peptide of urogenital tissues.]. Levofloxacin possesses a favorable pharmacokinetic profile with wide distribution into tissues and macrophages, good oral absorption, and a long biological half-life that allows for once daily dosing [34Piscitelli S.C. Spooner K. Baird B. Chow A.T. Fowler C.L. Williams R.R. et al.Pharmacokinetics and safety of high-dose and extended-interval regimens of levofloxacin in human immunodeficiency virus-infected patients., 36Chien S.C. Rogge M.C. Gisclon L.G. Curtin C. Wong F. Natarajan J. et al.Phamacokinetic profile of levofloxacin following once-daily 500mg oral and intravenous doses.]. No adverse effect related to LVFX was encountered.The role of Gram+ and Gram- bacteria in negative female and male fertility outcome has been reported [37Fujita Y. Mihara T. Okazaki T. Shitanaka M. Kushino R. Ikeda C. et al.Toll-like receptors (TLR) 2 and 4 on human sperm recognize bacterial endotoxins and mediate apoptosis., 38KAMIYAMA S. TERUYA Y. NOHARA M. KANAZAWA K. Impact of detection of bacterial endotoxin in menstrual effluent on the pregnancy rate in in vitro fertilization and embryo transfer.]. These findings indicate that IUMC may be involved in female infertility and may contribute to poor fertility outcome in women with endometriosis. These detrimental effects of IUMC and chronic endometritis on fertility outcome may be alleviated by single or combined treatment with LVFX and GnRHa. The decreased IUMC and concurrent decreased occurrence of chronic endometritis in response to LVFX or LVFX + GnRHa treatment as we found in our current study may improve fertility outcome in women who are suffering from endometriosis-related infertility.The occurrence of sub-clinical infection does not only occur in intrauterine environment but also in the cystic fluid of ovarian endometriosis [[11]Khan K.N. Fujishita A. Masumoto H. Muto H. Kitajima M. Masuzaki H. et al.Molecular detection of intrauterine microbial colonization in women with endometriosis.]. Although we did not analyze bacterial genera in the cystic fluid of ovarian endometrioma in our current study, we found that treatment with GnRHa, or LVFX, and GnRHa + LVFX was able to significantly decrease immunostaining levels of CD68 (inflammatory response), Ki-67 (cell proliferation), and CD31 (micro-vessels) in the biopsy specimens derived from the cyst walls of ovarian endometrioma. These findings were coincided with histology-proven improvement of ovarian endometrioma. In fact, we found that comparing to untreated group, both GnRHa and LVFX treatment caused either of disappearance, fragmentation or decreased thickness of epithelial lining cells as well as stromal cells population in the cyst walls (Fig. 8 and Fig. 9). These findings were associated with variable changes in the distribution of immune cells and tissue bleeding in cyst walls (Table 6). Further study is warranted to confirm these histological findings by time-dependent image analysis after each treatment.

There are some limitations in our current study: (1) This is a non-randomized observational study and prospective follow-up was not done. (2) We used single dose of LVFX as an initial observational study with the preference and consent of patients. In clinical practice, antibiotic treatment for a period of 10-14 days is recommended for the treatment of diagnosed CE in order to improve the fertility outcome. Our aim was not to treat CE by long time antibiotic but to examine the effect of a single dose of LVFX, if any, on IUMC and any change in the occurrence of CE. In fact, it is difficult to persuade patients to accept multiple doses of LVFX without indication. Our current findings could be a good reference for future study. (3) We cannot exclude the bias of any previous treatment with either immunosuppressing agent or antimicrobial agents. Unfortunately, we could not record the history of these treatments for our enrolled patients. (4) Our samples size was small in each treatment group of women with and without endometriosis. Further studies with large sample size are needed to confirm the validity of our current findings and their clinical application in women with endometriosis. (5) We could not collect biopsy samples of ovarian endometrioma from GnRHa + LVFX-treated women. Future study may address these unclear issues.

In conclusion, our current findings suggested that clinical application of a broad-spectrum antibiotic (LVFX) either alone or in combination with GnRHa may be effective in improving subclinical uterine infection with consequent decrease in the occurrence of chronic endometritis. The decreased uterine infection in response to LVFX and/or GnRHa was coincided with decrease in tissue inflammation, cell proliferation, and angiogenesis in endometria and endometriotic lesions with histology-proven improvement of ovarian endometrioma. In the field of reproductive medicine, we are still facing a big problem with the occurrence and recurrence of endometriosis and related pain symptoms after successful hormonal and surgical treatment. In this critical standpoint, clinical application of a widely available broad-spectrum antibiotic together with estrogen suppressing agent may ameliorate the risk of uterine infection, occurrence or recurrence of lesion, and related pain symptoms in women who are suffering from this hazardous disease. Future randomized prospective follow-up study with large sample size may address this issue and strengthen the validity of our current findings for clinical application and to aid in the development of personalized treatment strategy in women with endometriosis.

Acknowledgements

We thank Ms. Ayumi Tanaka and Ms. Yunhwa Lee of Kyoto Prefectural University of Medicine for their excellent technical assistance. The authors also thank all doctors of the Department of Gynecology and Department of Pathology, Saiseikai Nagasaki Hospital, Nagasaki, Japan for their kind and excellent assistance in collecting biopsy samples and sample preparation for this study.

Financial support: This work was supported in part by Grants-in-aid for Scientific Research (Grant Nos. 24592474, 15K10675, 18K09268 to KNK) from the Japan Society for the Promotion of Science (JSPS).

Conflict of interest: The authors declare that there is no financial or other conflict of interest related to this article.

Conflict of interest statement: The authors declare that there is no conflict of interest related to this article.