Dietary fish oil supplementation inhibits formation of endometriosis-associated adhesions in a chimeric mouse model

Using an established model of experimental endometriosis, we reported that adhesions rarely develop in mice that have been allowed to fully recover from ovariectomy (≥5 days) before intraperitoneal injection of human endometrial tissue, whereas adhesions are common among mice in which experimental endometriosis is established within 36 hours of surgery ( 4 ). Herein, we provided mice standard mouse chow or the same diet supplemented with 5% or 10% menhaden fish oil for 2 weeks before surgery and continuing until we killed them 5 days after establishment of experimental endometriosis. At necropsy, mice on the standard diet undergoing ovariectomy within 16 hours of receiving human endometrial tissues developed extensive adhesive disease associated with experimental endometriosis. Similar to our previous observations ( 4 ), all control mice developed lesions and adhesions at the site of surgical injury as well as at sites distal to the injury ( Table 1 ; Fig. 1A and B ). In contrast, both the adhesion score and experimental disease burden were significantly reduced in mice maintained on the 10% fish oil–supplemented diet ( Table 1 ; Fig. 1C and D ). Notably, the surgical site was the primary location for adhesion development in the supplemented mice, with distal occurrence rare. Recognizing that supplementation with 10% fish oil would be difficult to achieve in a patient population, we additionally examined the effectiveness of a 5% fish oil–supplemented diet. As shown in Table 1 and Figures 1E and 1F , we noted a reduction in both adhesion score and experimental disease burden similar to that observed in the mice provided 10% fish oil. As reflected by the adhesion scores, gross observations at necropsy suggested the adhesions in mice maintained on either fish oil diet were physically weaker and had fewer attachment points compared with those developing in control animals. To more fully characterize the nature of the adhesions, formalin-fixed paraffin embedded tissues were subjected to Masson trichrome and PTAH staining. As shown in Figure 2 , Masson trichrome staining of tissues from mice maintained on the standard diet revealed an intense bright blue stain ( Fig. 2A ) whereas adhesions/lesions obtained from the mice provided the fish oil–supplemented diets were only weakly stained ( Fig. 2B and C ). Because staining intensity has been demonstrated to be reflective of the abundance of collagen ( 29 ), these observations suggest that fish oil supplementation was associated with reduced collagen deposition. These data are consistent with our gross observations at necropsy, in which the adhesions present in mice receiving the standard diet appeared to be more substantial and had more attachment points compared with those present within the fish oil groups. Interestingly, PTAH histochemistry, which stains fibrin fibers dark blue, revealed minimal differences between control tissues and tissues from mice on either of the supplemented diets ( Fig. 2D–2F ). Adhesion formation has been associated with excessive immune cell influx at sites of wound healing ( 30 ); therefore, we assessed the presence of leukocytes within adhesions/lesions excised from mice provided the standard or fish oil–supplemented diet. Immunohistochemical localization of immune cells revealed a robust leukocyte infiltrate within the majority of both endometriotic-like lesions and adhesions obtained from control mice ( Fig. 3A and B ; 149.2 ± 18.6/×200 HPF; 91.9 ± 24.9/×400 HPF). In contrast, the presence of leukocytes within the adhesions/lesions removed from mice receiving the 10% fish oil–supplemented diet was consistently lower compared with the control group ( Fig. 3C and D ; 54.8 ± 13.8/×200 HPF; 30.6 ± 17.8/×400 HPF; P <.001). Although we did not have an adequate number of ×200 fields containing an appropriate amount of tissue to semiquantitatively assess the number of leukocytes present within the adhesions/lesions removed from the 5% fish oil–supplemented mice, microscopic examination and semiquantitative assessment at ×400 revealed a number of CD45+ immune cells similar to those found in the 10% fish oil–supplemented group ( Fig. 3E and F ; 27.3 ± 15.1/×400 HPF; P <.001). These data, taken together with our studies presented above, indicate that nutritional intervention with the omega-3 fatty acids, compounds with well known antiinflammatory actions, led to a reduction in leukocyte migration to ectopic sites which was also associated with a change in the physical nature of the adhesions.

Development of adhesive disease is a common postsurgical complication, with up to 50% of patients who undergo abdominal surgery developing adhesions ( 31 ). Appropriate wound healing after surgery requires both fibrin deposition to initially close the wound followed by fibrinolysis to break down the fibrin bands as healing progresses. However, if fibrinolysis is compromised, macrophages, fibroblasts, and endothelial cells may penetrate into the adhesion, leading to collagen deposition and the formation of a permanent fibrous adhesion ( 6 ). Importantly, a failure to appropriately resolve inflammation associated with wound healing can lead to chronic inflammation which further promotes postsurgical development of adhesive disease ( 9 , 10 , 32 , 33 ). Notably, surgical resection of endometriosis is a common procedure, placing these women at risk of developing postsurgical adhesions. Additionally, the chronic peritoneal inflammation associated with ectopic endometrial growth likely places endometriosis patients at an increased risk for the development of adhesions even in the absence of surgery ( 6 , 9 , 10 , 34 ). Therefore, there is an urgent need to identify therapeutic regimens that will reduce the risk of endometriosis-associated adhesive disease in the presence or absence of surgery. To experimentally address the relationship between endometriosis-related inflammation and adhesion formation, we previously described an animal model in which spontaneous adhesive disease develops following the introduction of human endometrial tissue fragments subsequent to peritoneal surgery to remove the ovaries. Although ovariectomy is a surgical component of our established experimental endometriosis model ( 35 ), introduction of human endometrial tissues into the peritoneal cavity after the wound healing process is complete does not result in adhesion formation. However, a time-course study revealed that peritoneal adhesions were common if initiation of experimental endometriosis occured within 36 hours of ovariectomy. Furthermore, the most significant adhesive disease was observed when human tissues were injected at 16 hours after surgery ( 4 ). Importantly, all mice were killed 5 days after tissue injection, providing ample time for establishment of experimental endometriosis and/or the development of adhesions. Within the peritoneal cavity, the innate immune system normally acts to promote clearance of displaced endometrial tissues following retrograde menstruation. However, under the additional influence of surgical injury, various cytokines and chemokines secreted by immune cells to promote wound healing may actually contribute to inflammation-associated establishment of both experimental endometriosis and adhesive disease ( 3 , 4 ). Importantly, temporal separation of the initiation of experimental endometriosis and surgical injury by >36 hours largely eliminated the occurrence of adhesions. These data suggest that therapeutic intervention directed at reducing peritoneal inflammation related to wound healing should limit the formation of adhesions in our experimental endometriosis model. Indeed, herein we demonstrated that providing mice with an “antiinflammatory” diet containing fish oil significantly reduced experimental endometriosis-associated adhesive disease. The present study also suggests that leukocyte infiltration may play a role in promoting dense collagen-rich adhesions as well as more extensive experimental disease. These findings are consistent with human studies demonstrating that endometriosis-associated adhesions exhibit greater numbers of inflammatory cells compared with infection-related adhesions ( 30 ). Taken together, the findings described herein support a prominent role of excess inflammation, involving the migration of leukocytes, in the development of adhesions associated with establishment of experimental endometriosis. However, dietary supplementation of mice before surgery and induction of experimental disease reduced leukocyte influx and collagen deposition. In contrast, fibrin deposition, which precedes collagen deposition by 2–3 days, was similar across all groups, suggesting that fish oil supplementation did not impede normal wound healing. Although nutritional intervention for adhesion prevention has received little attention, multiple studies suggest that an imbalance in the serum omega-6/omega-3 ratio can be associated with chronic inflammatory diseases ( 36 , 37 ). This is not surprising, because studies support a role of omega-3 fatty acids in promoting the resolution of inflammation associated with wound healing ( 38 ). Specifically, the omega-3 fatty acids serve as the precursors for production of the resolvins and protectins ( 36 , 39 ), endogenous molecules that have both antiinflammatory and proresolution effects ( 38 ). As described above, it is the failure of the resolution process that ultimately leads to the formation of permanent adhesions after injury. Unfortunately, humans and other mammals have only a limited capacity to synthesize the essential long-chain omega-3 fatty acids (specifically, eicosapentaenoic acid [EPA] and decosahexaenoic acid); therefore, these compounds must be obtained exogenously ( 40 ). At least one study suggests that an imbalance between serum levels of EPA and arachidonic acid correlate with the severity of endometriosis ( 41 ). However, to our knowledge, no study has examined the omega-6/omega-3 ratio regarding a patient’s susceptibility to developing adhesive disease. Despite the ample data that indicate a significant role of inflammatory processes in development of adhesive disease ( 4 – 6 , 30 ), at present, preventive measures center on the use of various pharmacologic agents and barrier materials. Our data suggest that an antiinflammatory diet that includes fish oil supplementation may be a beneficial adjuvant to reduce development of adhesions in women undergoing surgical treatment of endometriosis.

Approval for human tissue use was obtained from the Vanderbilt University Institutional Review Board and Committee for the Protection of Human Subjects. After written informed consent, endometrial samples (n = 9) were obtained by Pipelle biopsy (Unimar) during the proliferative phase (days 9–12) from a donor population (age 21–45 y) at Vanderbilt University Medical Center. Donors exhibited normal menstrual cycles with no known history of adhesions or endometrial disease, including endometriosis. A serum P level of <1.5 ng/mL at the time of biopsy was required for inclusion. Individuals with a history (<3 months) of hormonal therapy (i.e., oral contraceptives) or other medications that could affect study results (e.g., Omacor, Celebrex) were excluded. Endometrial specimens were immediately rinsed in Dulbecco Modified Eagle Medium/Nutrient Mixture F-12 (Gibco) to remove residual blood and mucous. Tissues were minced into 1–2-mm 3 cubes and maintained overnight under serum-free conditions in the same medium supplemented with 1 nmol/L E 2 (Sigma), 1% insulin-transferrin-selenium (Collaborative Biomedical), and 0.1% Excyte (Miles Scientific) and incubated in a 5% CO 2 humidified chamber at 37°C. Virgin, female (5 weeks) Athymic Nude- Foxn1 nu mice were purchased from Harlan Laboratories (Indianapolis) and housed in Vanderbilt University’s mouse vivarium under standard conditions (22°C and on a 12-h light-dark cycle) and provided food and water ad libitum. After 1 week of acclimatization, mice were randomized to receive low-phytoestrogen rodent chow (V502; Purina) or the same diet supplemented with 5% or 10% menhaden fish oil. The supplemented diets were produced on our behalf by Purina Test Diets and provided in pellet form similar to the control diet. Menhaden fish oil, donated by Omega Protein, has an established fatty acid profile (~40% omega-3 fatty acids) and was subjected to molecular distillation by that company to remove dioxins/PCBs. The fish oil diets were maintained in vacuum-sealed bags at −20°C until use and once provided to mice were replaced every 3 days. Mice were maintained on the fish oil–supplemented diet for ≥2 weeks before surgery and until we killed them (5 days after human tissue injection). All mice were weighed at the initiation of these studies (when mice were randomized to the various diets), at time of surgery, and when we killed them. At each time, weights were similar between groups, suggesting minimal differences in food consumption. An ideal diet has been suggested to contain a 3:1 ratio of omega-6 to omega-3; however, among Western countries the typical diet contains 15:1 omega-6 to omega-3 with ratios in the U.S. often reaching 40:1 ( 22 , 23 ). The diets used herein contained omega-6/omega-3 ratios of 10:1 (standard diet), 3:1 (5% fish oil diet), and 1.5:1 (10% fish oil diet). Protein, total fat, and energy content were similar among all of the diets. Mice were anesthetized with isoflurane and subjected to standard surgical ovariectomy via a single 5 mm dorsal/ventral incision between the rib cage and hind limb as previously described ( 4 ). At the time of surgery, all mice were implanted subcutaneously with a slow-release E 2 capsule assembled in our laboratory ( 24 ). Experiments described herein were approved by Vanderbilt University’s Institutional Animal Care and Use Committee in accordance with the Animal Welfare Act. We recently reported that endometriosis-associated adhesive disease develops in mice as a consequence of introduction of human endometrial tissue fragments near the time of peritoneal surgery ( 4 ). Specifically, at necropsy 5 days after tissue injection, we found that mice that received endometrial tissues 16 hours after ovariectomy exhibited extensive adhesive disease and robust experimental endometriosis ( 4 ). Therefore, in the present study, ovariectomy was performed 16 hours before intraperitoneal injection of human endometrial tissues ( Supplemental Fig. 1 , available online at www.fertstert.org ). After overnight culture, endometrial fragments were washed in sterile prewarmed phosphate-buffered saline solution (PBS) and injected intraperitoneally along the ventral midline with the use of an 18-gauge needle as previously described ( 25 ). Each biopsy was typically adequate for induction of experimental disease in 6–8 mice. Each experiment used one biopsy from a single donor and included both the control (standard diet) group of mice and either the 5% or 10% fish oil–supplemented group or both experimental groups. Within each separate experiment (using a single biopsy), 2–3 mice per treatment group were established with experimental disease. Each group was replicated with at least four human endometrial biopsies. At necropsy, 5 days after intraperitoneal introduction of human tissues, the presence or absence of peritoneal adhesions and the presence and extent of ectopic sites of endometrial growth was determined. Scoring of adhesions was based on a system described by Lauder et al. ( 26 ), which considers the number, strength, and distribution of adhesions. Specifically: 0 = no adhesions; 1 = thin filmy adhesions; 2 = more than one thin adhesion; 3 = thick adhesion with focal point; 4 = thick adhesion with planar attachment; 5 = very thick vascularized adhesions or more than one planar adhesion. Endometriotic-like lesion burden was scored by number and size, with size measured in two dimensions, the larger denoted “a” and the smaller denoted “b.” The total volume was calculated by standard methodology ( 27 ) with the formula: V = a × b 2 × 0.5. All tissues excised from mice at necropsy were formalin fixed, paraffin embedded, and subjected to standard hematoxylin and eosin staining. Additional sections were further subjected to Masson trichrome staining to assess the extent of collagen deposition and phosphotungstic acid hematoxylin staining (PTAH) to assess fibrin content. These stains were performed by the Vanderbilt Translational Pathology Shared Resource Core Laboratory with the use of standard methods. Immunohistochemical localization of CD45, a panleukocyte marker, was used to assess the infiltration of leukocytes within adhesions/lesions. Briefly, formalin-fixed paraffin-embedded tissue sections were deparaffinized and subjected to antigen retrieval with the use of 1× Citrate Retrieval Buffer (Biogenex) in deionized water (dH 2 0) at 70–75°C for 25 minutes. After cooling to room temperature, sections were rinsed twice with dH 2 0 for 5 minutes each followed by blocking of endogenous peroxidase activity with the use of cold 3% hydrogen peroxide in methanol for 20 minutes. Sections were again rinsed twice with dH 2 0 for 5 minutes each followed by incubation with 1× PowerBlock (Biogenex Laboratories) in dH 2 0 for 15 minutes at room temperature. Sections were rinsed twice with 1× PBS for 5 minutes each then incubated at 4°C overnight with the primary antibody (1:150 rat antimouse CD45; BD Pharmagen) diluted in 1× PBS containing 0.5% Triton-X. Following two washes in 1× PBS, sections were incubated with an antirat secondary antibody (Biogenex Laboratories) at room temperature for 20 minutes, followed by two PBS washes for 5 minutes each. Sections were then incubated at room temperature using Strepavidin Peroxidase Label (Biogenex Laboratories) for 20 minutes, followed by two washes in 1× PBS and then chromagen detection with the use of Impact DAB (Vector Laboratories or Nova Red) according to the manufacturer’s instructions. Finally, sections were counterstained in Gill hematoxylin for 2 minutes followed by running tap water for 6 minutes, then dehydrated and coverslipped with the use of Cytoseal XYL (Thermofisher Scientific). Leukocyte infiltrates within adhesions/lesions were enumerated by two authors (D.R.G., K.B.T.) in five random high-power fields (HPF) per group (at ×200 and at ×400 magnification) with the use of a method similar to that described by others for assessment of immune cell populations in tumors ( 28 ). The same five fields from each group and each magnification were enumerated by both observers, and the resulting ten numbers were averaged to semiquantitatively determine the number of CD45-positive cells present within control and 10% fish oil mice. Semiquantitative analysis of leukocyte infiltrates in mice receiving the 5% fish oil diet was limited to ×400 magnification owing to the limited number of adequately sized samples available. All slides were viewed with an Olympus BX51 microscope system and images captured with an Olympus DP71 digital camera. Analyses were performed with Graphpad Prism-5 software and presented as mean ± SEM. The statistical difference between samples was determined using one-way analysis of variance followed by Tukey post hoc test. P <.05 was considered to be significant.