Periovulatory Leukocyte Infiltration in the Rat Ovary

Abstract

Ovulation is preceded by intraovarian inflammatory reactions that occur in response to the preovulatory gonadotropin surge. As a main inflammatory event, leukocytes infiltrate the ovary and release proteolytic enzymes that degrade the extracellular matrix weakening the follicular wall, a required step for follicle rupture. This study aimed to quantitatively measure the infiltrating leukocytes, determine their cell types, and localize infiltration sites in the periovulatory rat ovary. Cycling adult and gonadotropin-stimulated immature rats were used as animal models. Ovaries were collected at five different stages of estrous cycle in the adult rats (diestrus, 1700 h; proestrus, 1500 h; proestrus, 2400 h; estrus, 0600 h; and metestrus, 1700 h) and at five different time points after superovulation induction in the immature rats (pregnant mare’s serum gonadotrophin, 0 h; pregnant mare’s serum gonadotrophin, 48 h; human chorionic gonadotropin, 6 h; human chorionic gonadotropin, 12 h; and human chorionic gonadotropin, 24 h). The ovaries were either dissociated into a single cell suspension for flow cytometric analysis or fixed for immunohistochemical localization of the leukocytes. Similar numbers of leukocytes were seen throughout the estrous cycle (∼500,000/ovary), except proestrus 2400 when 2-fold higher numbers of leukocytes were found (∼1.1 million/ovary). A similar trend of periovulatory rise of leukocyte numbers was seen in the superovulation-induced immature rat model, recapitulating a dramatic increase in leukocyte numbers upon gonadotropin stimulation. Both macrophage/granulocytes and lymphocytes were among the infiltrating leukocytes and were localized in the theca and interstitial tissues, where platelet-endothelial cell adhesion molecule-1 and intercellular adhesion molecule-1 may play roles in the transmigration of leukocytes, because their expressions correlates spatiotemporally with the infiltrating leukocytes. In addition, a strong inverse relationship between leukocyte numbers in the ovary and spleen, as well as significant reduction of leukocyte infiltration in the splenectomized rats, were seen, indicating that the spleen may serve as an immediate supplier of leukocytes to the periovulatory ovary.

A comprehensive quantitative measurement of leukocyte infiltration in the ovaries of adult and superovulation-induced immature rats is presented.

The ovulatory process, ending with rupture of the follicular wall and expulsion of the oocyte, is a central event in the reproductive cycle. Up to the middle of the 1960s, it had been generally accepted that increased intrafollicular pressure was the driving force of follicle rupture (1). However, the consensus was challenged by reports that there is no significant increase in intrafollicular pressure before the rupture of the follicle and that artificially increased intrafollicular pressure induced by an injection of saline into rabbit follicles does not induce follicle rupture (2,3). These reports steered researchers to view the “decreasing tensile strength of the follicular wall” as a main causal factor in follicle rupture, which inevitably emphasized the importance of proteolytic enzymes and the plasminogen system in ovulation. Since then, numerous studies have implicated matrix metalloproteinases and the plasminogen activators in the degradation of the collagenous tissues that are present in the basement membrane, theca externa, and tunica albuguinea of the follicular wall (4,5). Recent studies have indicated that follicular contraction may serve as a contributing factor in the follicle rupture (6,7,8,9).

Both ovarian cells and infiltrating leukocytes secrete proteases in the preovulatory ovary (4,10,11). However, the mechanism of leukocyte infiltration is yet to be established. The present study aimed to quantitatively measure the infiltrating leukocytes, determine their cell types, and localize infiltration sites in the periovulatory rat ovaries. In addition, an attempt was made to determine whether the spleen serves as an additional source of the infiltrating leukocytes to the ovary. Leukocytes are unique in that they can traverse to many parts of the body where and when needed. Specific cell adhesion molecules expressed on inflamed tissues attract leukocytes. The main function of these leukocytes is to release chemokines and cytokines that directly or indirectly signal the further recruitment of leukocytes, as well as modulate proteolytic activity. Supporting this idea, a number of molecules that are commonly associated with immunological responses have been found to be present in the ovary, including various cytokines and chemokines (12,13). Changes in these molecules are associated with the cyclic changes in estrous cycle compounding their role in the ovulatory process (14). Here, we report a comprehensive quantitative measurement of leukocyte infiltration in the ovaries of adult and superovulation-induced immature rats.

Materials and Methods

Animal handling/hormone injection and estrous cycle determination

Four-month-old virgin adult, 23- to 24-d-old immature, and 6-wk-old splenectomized Sprague Dawley rats were purchased from Harlan (Indianapolis, IN) and used. They were either subjected to estrous cycle determination by vaginal cytology for a minimum of 2 wk (adult rats) or superovulated by pregnant mare’s serum gonadotrophin (PMSG) and human chorionic gonadotropin (hCG) injection (immature rats) before collecting ovaries, spleens, and peripheral blood samples at five different time points, followed by flow cytometric cell counting using one ovary from each rat and immunohistochemistry for localizing leukocytes using the other ovary. Estrous cycles of the adult rats were determined as previously described (15,16) by vaginal lavage techniques. Briefly, vaginal lavage was performed by flushing the vagina with 0.9% sodium chloride daily in the morning at the same time each day. The flushed vaginal fluid was examined microscopically and scored. The cycling data were expressed as estrus, metestrus, diestrus, or proestrus. Estrus was determined by the presence of cornified cells. Metestrus was determined by the presence of large round cells with an irregular border. Diestrus was scored by the presence of a high density of leukocytes, whereas proestrus was indicated by the presence of small nucleated cells. Superovulation was induced in immature rats by ip injection with 10 IU of PMSG (G4877; Sigma, St. Louis, MO) followed 48 h later by 10 IU of hCG (CG10; Sigma). Euthanasia was performed by CO₂ asphyxia followed by cervical dislocation at the designated times. For adult rats, five different stages of estrous cycle were selected: diestrus 1700 h (D1700), proestrus 1500 h (P1500), proestrus 2400 h (P2400), estrus 0600 h (E0600), and metestrus 1700 h (M1700). For immature rats, the selected time points were PMSG 0 h (PMSG 0), PMSG 48 h, hCG 6 h (hCG 6), hCG 12, and hCG 24. All the animal procedures were approved by the University of Kentucky Animal Care and Use Committee.

Isolation and flow cytometric analysis

Spleen

At the designated times, animals were killed by CO₂ inhalation followed by cervical dislocation. The spleens were harvested and single cell suspensions of splenocytes prepared and counted as previously described (17).

Ovaries

Ovaries were collected at the designated times. Single cell suspensions of ovaries were prepared by modifying a method previously described for lung tissues (18). Briefly, ovaries were removed, placed into a small Petri dish containing 2 ml of PBS, and cut into 15–20 pieces using a fine blade (Accu-Edge 4689; Sakura, Tokyo, Japan). The tissues were then transferred using a Pasteur pipette to a 15-ml tube containing 10 ml of PBS. Debris was removed from the preparation by allowing the tissues to settle to the bottom of the tube, and all but 2 ml of the supernatant was removed; 2 ml of collagenase digestion solution containing 3.5 U of collagenase type I (17100-017; Invitrogen, Carlsbad, CA), 1000 U of deoxyribonuclease I (D4527; Sigma), and 40 mg of BSA (017K0723; Sigma) in H-199 media (12350-039; Life Technologies, Inc., Carlsbad, CA) was added to the tissue suspension, and placed in a water bath at 37 C for 30 min. After incubation, tissues were further dissociated by repeatedly passing them through an 18-G needle attached to a 5-ml syringe. Then, the digested tissue was filtered through a 40-μm filter into a 50-ml tube, washing the filter with an additional 10 ml of PBS. The cell suspension was centrifuged at 250 × g for 5 min and resuspended in 2 ml of PBS. The tissue digest that failed to pass the filter was not included in the flow cytometry assay and therefore was unaccounted for in the presented data. For flow cytometry (FCM) analysis, cells (1 × 10⁶ cells) were stained with fluorochrome-conjugated monoclonal antibodies against flurochrome-conjugated antibodies specific for CD45-PE-Cy5 (554839; BD Biosciences, San Jose, CA), CD11b/c-FITC (554861; BD Biosciences), RP-1-PE (550002; BD Biosciences), and CD3-FITC (554832; BD Biosciences) for 20 min on ice (specific antibody information is described elsewhere in the Materials and Methods). Cells were washed and fixed in 2% formaldehyde. After washing and resuspension in PBS, cells were analyzed using FACSCalibur Cytofluorometer (BD Biosciences, San Jose, CA) as previously described (17). Data analysis was performed using Winlist 6.0 3D software (Verity Software House, Topsham, ME).

Blood collection and complete blood count analysis

The blood was collected by cardiac puncture using 15-ml syringes (Falcon; BD Biosciences). After collection, 1–3 ml of whole blood was immediately placed into EDTA tubes (Falcon; BD Biosciences) and mixed by inversion to prevent clotting. At the time of analysis, all blood samples were mixed gently by inversion and run on a hematology analyzer (Hemavet 950 FS; Drew Scientific, Inc., Dallas, TX) to obtain complete blood count with differential analysis (% neutrophils, % lymphocytes, % monocytes, % eosinophils, and % basophils). For serum collection, the extra blood collected from cardiac puncture was immediately placed into Eppendorf tubes and allowed to clot completely for 10–15 min. Blood was then centrifuged at 3500 × g for 20–25 min using Eppendorf 5415-D microcentrifuge (Eppendorf, Hamburg, Germany), and serum was removed and placed into new Eppendorf tubes and frozen at −80 C until the time of analysis.

Immunohistochemistry

Serial frozen sections of 7-μm thickness were washed in PBS and placed in 3% H₂O₂ for 15 min to remove endogenous peroxidase activity, followed by three washes with PBS and blocked with 3% albumin solution from bovine serum (A-9576; Sigma) for 20 min at room temperature. Sections were then washed three times in PBS and incubated with mouse monoclonal antirat CD3 (550295; BD Biosciences), mouse monoclonal antirat CD11b/c (550299; BD Biosciences), mouse monoclonal antirat CD31 (M550300; BD Biosciences), and mouse monoclonal antirat CD45 (550566; BD Biosciences) antibodies and incubated overnight at 4 C in a humidified chamber. Sections were washed three times in PBS and then incubated with biotin goat antimouse Ig (550337; BD Biosciences), biotin rat antimouse IgG2a (550333; BD Biosciences), and biotin rat antimouse IgG3 (553401; BD Biosciences) secondary antibodies for 1 h at room temperature. After three washes with PBS, the antigen-antibody complexes were detected using a streptavidin horseradish peroxidase (550946; BD Biosciences) for 30 min followed by three rinses in PBS. The resulting signal was developed with diaminobenzidine (D-4293; Sigma), and sections were counterstained with Mayer’s Hematoxylin (S3309; Dako North America, Inc., Camarillo, CA) and mounted with glycerol (G-5516; Sigma) on glass slides. The same procedure omitting primary antibody was used for control groups. Photomicrographs were taken with an Olympus microscope using Spot Imaging software version 2.1B (BX5ITF; Diagnostic Instruments Inc., Sterling Heights, MI). All samples for each individual antibody were exposed to the same protocol at the same time and were stained using the same incubation periods.

Statistical analysis

When applicable, results were subjected to statistical analysis. Data were analyzed and plotted using SigmaPlot version 10.0 (Systat Software, Inc., Chicago, IL). On graphs, error bars represent one ±sem around the average of data per group. To determine the statistical significance between groups, ANOVA was performed followed by post hoc analysis using Dunnett’s method for the determination of individual differences. Data that failed normality testing was normalized using log transformation. The difference was considered significant when P ≤ 0.05.

Results

Leukocytes in the adult ovary

Ovaries from five different estrous cycle stages were examined: D1700, P1500, P2400, E0600, and M1700. In our animal facility, LH surge occurs in the early evening of proestrus (P1800–P1900). Therefore, P2400 was about 5–6 h after LH surge had occurred. The total numbers of ovarian cells isolated dramatically increased by 2 million between P1500 and P2400, and then, the numbers decreased back to basal level by E0600 (Fig. 1). Leukocytes (CD45⁺ cells) accounted for approximately one quarter (∼500,000 cells) of all cells isolated from a P2400 ovary. We further characterized the populations of leukocytes present in the ovary using a combination of fluorescent-conjugated antibodies to leukocyte surface antigens. Lymphocytes were identified using antibodies to the CD3 cell surface marker, and macrophage/granulocyte cells were identified as staining double positive to CD45⁺ and CD11b/c⁺ antibodies. This analysis revealed that both CD11b/c⁺ and CD3⁺ cells contributed to this increase in numbers. Interestingly, although the greatest increase in numbers of ovarian leukocytes was observed at P2400, the greatest increase in the proportion of leukocytes was seen at E0600 (Figs. 1 and 2).

Figure 1.

Ovarian leukocyte evaluation in naturally cycling adult SD rats. Tissue samples were collected at the respective time points. Cells isolated from the ovaries of rats were stained with a fluorescent-conjugated antibodies specific for the leukocyte marker, CD45, CD11b/c, and CD3. Stained cells were analyzed on a FACSCalibur flow cytometer. Actual numbers of the cells counted (left graphs) and percentages of the cells (right graphs) are shown. Mean ± sem of n = 3 rats is shown, except D1700 and M1700, where n = 6. Values with different superscript letters differ (P < 0.05, ANOVA + Dunnett).

Figure 2.

Representative leukocyte population patterns in the naturally cycling adult rats. Graphs were generated from the flow cytometry. The blue-colored areas represent CD45⁺ leukocyte population. Representative histogram data for an individual ovary is shown (upper panels). The numbers are percentages of leukocytes out of total numbers of ovarian cells counted (upper graphs). The other ovary of the same rat at each time point was used for histology by hematoxylin/eosin staining (lower panels).

Leukocytes in the immature ovary

The increase in leukocyte numbers in the evening of proestrus was consistent with previous reports that LH surge stimulated ovarian leukocyte infiltration (11). To gather more direct evidence that LH stimulation would induce the ovarian leukocyte infiltration, we repeated the experiment under superovulation regimen (Fig. 3). Immature rats (23–24 d of ages) were treated for synchronized folliculogenesis and ovulation by injecting 10 IU of PMSG followed by 10 IU of hCG injection 48 h after PMSG injection. Forty-five hours after PMSG injection (PMSG 45), hCG 6, hCG 12, and hCG 24 specimens were collected and processed the same way as they were for the adult cycling rats. Flow cytometric analysis revealed a pattern of increasing numbers of ovarian leukocytes after hCG administration (Fig. 3) similar to the rise of leukocyte numbers in the evening of proestrus in the adult rats (Fig. 3) but in a much more dramatic way (Figs. 3 and 4). Before PMSG injection, approximately 15,000 leukocytes (8.52%) resided in the ovary. After PMSG injection, there was no significant increase in ovarian leukocyte numbers. PMSG 45, about the same numbers of leukocytes were counted (∼15,000 leukocytes/ovary; 11.07% of total cell counted). Six hours after hCG injection, however, the leukocyte numbers sharply increased by 5-fold (∼75,000 leukocytes/ovary; 21.22% of total cells counted). After that, the percentage started to decrease; at hCG 12, the leukocyte numbers reached approximately 150,000 (17.11% of total cells counted), at which time ovulation would take place. The leukocyte numbers remained at this level until hCG 24.

Figure 3.

Ovarian leukocytes in superovulated immature rats. Cells isolated from the ovaries were stained with a fluorescent-conjugated antibody specific for the leukocyte marker, CD45. Stained cells were analyzed on a FACSCalibur flow cytometer. Representative histogram data for an individual ovary are shown (upper panels). Three rats were used in each time point. The numbers are percentages of leukocytes out of total numbers of ovarian cells counted (upper graphs). The other ovary of the same rat at each time point was used for histology by hematoxylin/eosin staining (lower panels).

Figure 4.

Analysis of leukocyte subpopulations within the ovary during stimulated ovulation in immature rats. Ovaries were collected from superovulated rats killed at the indicated times and stained with fluorescent-conjugated antibodies specific for leukocyte markers. Stained cells were analyzed by four-color analysis on a FACSCalibur flow cytometer. Graphs, Total numbers (left) and percentages (middle) of CD45, CD11b/c, and CD3 subsets as determined by FCM from ovaries of superovulated immature rats. The sem of n = 3 rats is shown. Values with different superscript letters differ (P < 0.05, ANOVA + Dunnett). Right panels, Immunohistochemical staining of tissue sections from ovaries at hCG12 stained with the indicated antibodies. GC, Granulosa cell layer; TH, theca cell layer. No positive staining was seen in the sections that were processed without primary antibody.

Double staining analysis with CD45+CD3+ and CD11b/c+ antibodies showed that both macrophage/granulocyte and lymphocyte populations infiltrate the ovary after hCG injection (Fig. 4). Interestingly, although the total number of lymphocytes increased in the ovary after hCG injection, the increase was not seen until hCG 12, and even at that time point, there was a substantial reduction in the proportion of CD3 cells present in the ovary (Fig. 4). Immuno histochemical assays found that leukocytes were mostly located in the theca and interstitial tissues and rarely inside follicles (Fig. 4C and Supplemental Fig. 1, published on The Endocrine Society’s Journals Online web site at http://endo.endojournals.org).

Preferential infiltration of agranular leukocytes

To further dissect the species of infiltrating leukocytes, the cells isolated from ovaries were simultaneously stained with anti-granulocyte (RP-1) and CD11b/c⁺ antibodies for flow cytometric analysis. Interestingly, we observed a preferential infiltration of CD45⁺ CD11b/c⁺ RP-1⁻ (RP-1⁻) cells over CD45⁺ CD11b/c⁺ RP-1⁺ (RP-1⁺) cells. Although initially RP-1⁺ cells outnumbered RP-1⁻ cells, gonadotropin injection dramatically shifted the ratio; the RP-1⁻ cells accounted for 1.2% (2800 cells) at PMSG 0, 2.47% (3400 cells) at PMSG 45, 5.1% (19,000 cells) at hCG 6, and 6.0% (85,000 cells) at hCG 12 of all cells counted (see legend to Fig. 5). During the same period of time, RP-1⁺ cells did not show any dramatic changes in numbers or proportion (2.0%, 3300 cells) of cells at PMSG 0, 0.9% (3200 cells) at hCG 6, and 1.0% (10,000 cells) at hCG 12. When the same cell counting strategy was applied for the splenocytes, a strong inverse relationship between ovarian and spleen RP-1⁻ cell numbers, as well as percentage, were evident. RP-1⁻ accounted for 21.3% (14.2 million cells) at PMSG 0, 21.34% (15.6 million cells) at PMSG 45, and after gonadotropin stimulation, 17.78% (5 million cells) at hCG 6, and 18.4% (7.2 million cells) at hCG 12. As was for ovary, during the same period of time, RP-1⁺ cells did not show any dramatic changes in numbers or proportion (0.82%, 540,000 cells) of cells at PMSG 0, 0.93% (250, 000 cells) at hCG 6, and 0.7% (260,000 cells) at hCG 12 in the spleen (Fig. 5).

Figure 5.

Monocyte/macrophage distribution in the ovary vs. spleen during induced ovulation. Immunophenotype of ovarian cells and splenocytes form superovulation treated rats. CD45⁺ cells were further stained for CD11b/c and RP-1 antibodies. The results represent mean ± sem of n = 3 rats is shown. An asterisk indicates a significant difference compared with PMSG 0 (P < 0.05, ANOVA + Dunnett).

Sources of leukocytes infiltrating the ovary

To determine whether the inverse relationship of RP-1⁻ cell counts (Fig. 5) was applicable to the total leukocyte numbers in the ovary and spleen during the period of assay, leukocyte subpopulations were enumerated in the superovulated rats. After hCG injection, during the period of time when the intraovarian leukocyte numbers increased, the leukocyte numbers in the spleen sharply decreased (Fig. 6). The same inverse relationship was also observed in the adult rats (Supplemental Fig. 1). Consistent with this finding, significantly less numbers of leukocytes infiltrated the ovaries of the splenectomized rats (Fig. 7). Interestingly, although there was a slight increase in total peripheral blood cell counts at P2400, this was not significant (see Supplemental material). Interestingly, although there was a clear trend of inverse relationship in the numbers of leukocytes in the ovary and spleen, the decreased numbers in the spleen was significantly more than the increased numbers of leukocytes in the ovary (Fig. 6), indicating that splenocytes may also infiltrate other tissues.

Figure 6.

Comparison of leukocytes numbers between ovary and spleen. Leukocyte numbers found in the ovary and spleen were counted under superovulation treatment. Mean ± sem of n = 3 rats is shown. An asterisk indicates a significant difference compared with PMSG 0 (P < 0.05, ANOVA + Dunnett).

Figure 7.

Analysis of leukocyte subpopulations within the ovary of splenectomized (SPLX) rats during stimulated ovulation. Ovaries were collected from superovulated 6-wk-old rats killed at the indicated times and stained with fluorescent-conjugated antibodies specific for leukocyte markers. Stained cells were analyzed by four-color analysis on a FACSCalibur flow cytometer. Total numbers of cells, CD45⁺, CD45⁺ CD3⁺, and CD45⁺ CD11b/c⁺ subsets as determined by FCM. Data are the mean ± sem of n = 3. Values with different superscript letters differ (P < 0.05).

Ovarian expression of modulators of leukocyte recruitment

Leukocyte recruitment requires expression of a cohort of cell adhesion molecules. We first examined the expression pattern of platelet-endothelial cell adhesion molecule-1 (PECAM-1) (CD31). PECAM-1 is expressed in the tight junction areas of endothelial cells, where it provides a key role in the transmigration of leukocytes. PECAM-1 is a member of the immunoglobulin superfamily, and signaling through its intracellular domain can, within seconds, stimulate the activation of other inflammatory genes such as intercellular adhesion molecule-1 (ICAM-1) (19). Immunohistochemistry showed a massive expression of PECAM-1 in the interstitial and theca layer of the ovary at hCG 12. We further examined the expression patterns of other cellular adhesion molecules using rat ovarian gene expression database (20). This provides immediate analysis of temporal gene expression profiles for over 28,000 genes in intact ovaries, granulosa cells, and residual ovarian tissues during follicular growth and the periovulatory period (20). Figure 8 shows mRNA expression patterns for three major cellular adhesion molecules commonly found in acute inflammatory responses: ICAM-1, vascular adhesion molecule-1 (VCAM-1), and E-selectin (Fig. 8). Interestingly, although ICAM-1 expression was increased by hCG injection, no dramatic change of mRNA expression was seen for VCAM-1 and E-selectin. Adhesion molecule expression and the selective expression of their ligands on leukocytes can determine a preferential infiltration of specific leukocyte subpopulations (21). Although ICAM-1 and VCAM-1 (and their corresponding ligands; lymphocyte function-associated antigen 1 and very late antigen-4) are involved in lymphocyte infiltration, monocyte infiltration is mostly reliant on ICAM-1:lymphocyte function-associated antigen 1 interactions (22). This adhesion molecule expression pattern (increased ICAM-1 at hCG 6) is consistent with our leukocyte subpopulation results (Fig. 4), where macrophage/granulocyte cells increase at hCG 6 in numbers and percentage, whereas lymphocytes (CD3) increased in numbers but decreased in percentage of total leukocytes.

Figure 8.

Leukocyte recruitment in the ovary. Left, Rat ovarian gene expression database profiles of mRNAs for the adhesion molecules VCAM-1, ICAM-1, and E-selectin during folliculogenesis and in the periovulatory period. Mean ± sem of n = 2 is shown. *, P < 0.05 (ANOVA + t test) compared with PMSG 48 h. Right, Immunostaining of hCG 12 ovary using anti-PECAM-1 antibody. aF, Antral follicle; GC, granulosa cells. No positive staining was seen in the sections that were processed without primary antibody.

Discussion

Because Espey (23) formally hypothesized ovulation as an acute inflammatory response to gonadotropin stimulation in 1980, the “dynamic” nature of ovulation has been a subject of great interest at all levels of modern biology from the study on a specific gene regulation to the systemic interactions between distal organs. Accordingly, as main mediators of ovarian inflammatory responses, leukocytes have become one of the major subjects of investigation (24,25,26,27). Leukocytes circulate in the blood, become attracted by adhesion molecules in inflamed tissues, and transverse the blood vessel wall infiltrating interstitial tissue to the site of their action. Cytokines that are initially released by the inflamed tissue play a crucial role in increasing adhesion molecule expression on endothelial cells and also up-regulating the corresponding receptor expression on leukocytes, together greatly enhancing leukocyte migration to the target tissues. Once recruited to the inflamed site, leukocytes release chemokines and cytokines that further accelerate recruitment and function of leukocytes. Eventually, however, the main leukocyte functionality is exerted by their proteolytic activity through releasing proteases. Supporting Espey’s hypothesis, a number of molecules, including various cytokines, chemokines, and proteases that are commonly associated with immunological responses are also present in the preovulatory ovary (12,13). Their ovarian levels fluctuate along the estrous cycles compounding their role in the ovulatory process (14). The present study reports systemic quantization of leukocyte infiltration in the ovary during estrous cycle and in an induced ovulatory period.

We were first surprised by the finding that over a half million leukocytes reside in an adult rat ovary at any moment and that another half million leukocytes infiltrate the ovary within hours after LH surge, increasing the total number of leukocytes to over a million in an ovary before ovulation occurs (Figs. 1 and 2). The very same trend of leukocyte infiltration was seen in the gonadotropin-injected immature rats except that there were fewer resident cells before hCG injection but a much greater rate of leukocyte infiltration than in the adult rats. Only approximately 15,000 leukocytes/ovary were found before hCG injection. However, in just 6 h, after hCG injection, the leukocyte numbers increased to approximately 75,000 leukocytes/ovary, a 5-fold increase. In the next 6 h, the leukocyte numbers doubled to approximately 150,000 leukocytes/ovary by hCG 12, when ovulation begins to occur. The leukocyte numbers at this stage account for 10% of total cells dissociated from the ovary (Figs. 3 and 4). Two findings are of particular interest. First, the numbers of leukocytes (residential and infiltrating) are enormous. Second, the infiltration occurs very rapidly within a few hours.

The large number of “resident” leukocytes detected in mature rats throughout estrus (Fig. 1) may be required for the rapid infiltration and function of immune cells for ovulation. Typical immune response relies on a cascade of events between leukocytes and endothelial cell adhesion molecules finalizing in transmigration across the endothelial layer, taking up to 7 d to accomplish (21,28). The resident leukocytes, however, are able to instantly respond to stimuli by up-regulating adhesion molecules to rapidly recruit leukocytes to the site of inflammation (29). Our data indicates that neutrophils (RP-1⁺) remain at a constant level in the ovary, constitute a high proportion of the resident leukocytes, and are not influenced by estrous cycles in rats. The majority of infiltrating leukocytes in the periovulatory period are of the RP-1⁻ phenotype (Fig. 5), indicating monocyte/macrophage. Neutrophils, in fact, have been well documented in ovarian function mostly due to their capacity to produce a plethora of proteases (30,31,32). Although our data are suggestive of monocyte modulation of resident neutrophils, further analysis of the function of both populations is needed to determine the precise role of each cell type in ovulation.

Inflammatory sites usually recruit leukocytes from both bone marrow and peripheral circulation, reducing circulating numbers. However, in our studies, at the time of rapid infiltration of leukocytes to the ovary, we observed an increase in peripheral white blood cell counts (Supplemental Fig. 2). Further analysis revealed a large proportion of these cells to be of the monocyte/macrophage lineage (RP1⁻ cells) (Supplemental Fig. 2). The more than 30-fold increase in the numbers of RP-1⁻ cells immediately before ovulation and the speed at which they infiltrate (Figs. 5 and 6) indicate that neither the peripheral blood nor bone marrow serves the source of these cells. Evaluation of splenocytes during ovulation (natural or stimulated) revealed a distinct losing pattern of them while the ovary gains leukocytes (Figs. 5 and 6 and Supplemental Figs. 2 and 3), raising the possibility of spleen as a provider of leukocytes for the ovary. We tested this possibility by repeating the leukocyte counting under superovulation protocol using splenectomized rats. In agreement with our hypothesis, significantly reduced leukocyte infiltration was observed in the spelectomized rats. To our knowledge, this is the first report of the spleen as a possible reservoir for periovulatory leukocyte infiltration and supports the hypothesis of gonadotropin-triggered rapid leukocyte infiltration in ovulation. Supporting this hypothesis, Swirski et al. (33) recently showed that the spleen serves as a reservoir for specific leukocyte populations in a mouse model of acute myocardial infarction.

As has been well documented (34,35,36), most leukocytes are localized in proximity to endothelial cells (Fig. 4 and Supplemental Fig. 1), indicating that infiltrating leukocytes stay close to the site of infiltration without migrating to distal tissues once they transverse blood vessel walls. In support of the idea, in no case was a leukocyte found in the granulosa cell layer (Supplemental Fig. 1) except in the follicles whose basement membrane was extensively disrupted (Fig. 4). This is an expected result, because the granulosa cell layer is not vascularized (no direct infiltration by a leukocyte) but separated by basement membrane from a theca layer that is highly vascularized. The basement membrane may act as a barrier of cell migration until the extracellular matrix of the basement membrane is broken down by the proteolytic enzymes produced by the leukocytes, as well as other ovarian cells. After extracellular matrix break down, leukocytes may further participate in corpus luteum formation (14,37,38).

Adhesion molecules are crucial in the recruitment of leukocytes from the peripheral blood into sites of inflammation (39). Here, we demonstrate that increased leukocyte infiltration corresponds to increased ICAM-1 and PECAM-1expression immediately before ovulation. ICAM-1 mediated leukocyte attachment to endothelial cells is crucial in the initial steps of leukocyte migration, whereas PECAM-1 expression in the tight junction area of endothelial cells actively supports the transmigration process (28). A high level of PECAM-1 protein was seen in the periovulatory ovary, and ICAM-1 mRNA expression was up-regulated by hCG injection (Fig. 7), further implicating their involvement in the periovulatory recruitment of leukocytes.

Taken together, the present study provides quantitative temporal profiles of infiltrating leukocytes in the ovary and identifies the spleen as a potential contributor to the influx of inflammatory cells.