- Research article
- Open Access
Immunomodulation in the canine endometrium by uteropathogenic Escherichia coli
- Sofia Henriques†1,
- Elisabete Silva†1,
- Marta F. Silva1,
- Sandra Carvalho2,
- Patrícia Diniz1,
- Luís Lopes-da-Costa1 and
- Luisa Mateus1Email authorView ORCID ID profile
© The Author(s) 2016
- Received: 2 March 2016
- Accepted: 14 September 2016
- Published: 9 November 2016
This study was designed to evaluate the role of E. coli α-hemolysin (HlyA) in the pathogenesis of canine pyometra, and on the immune response of canine endometrial epithelial and stromal cells. In Experiment 1, the clinical, hematological, biochemical and uterine histological characteristics of β-hemolytic and non-hemolytic E. coli pyometra bitches were compared. More (p < 0.05) metritis cases were observed in β-hemolytic E. coli pyometra uteri than in non-hemolytic E. coli pyometra uteri. β-hemolytic E. coli pyometra endometria had higher gene transcription of IL-1β and IL-8 and lower gene transcription of IL-6 than non-hemolytic E. coli pyometra endometria (p < 0.01). In Experiment 2, the immune response of endometrial epithelial and stromal cells, to hemolytic (Pyo18) and non-hemolytic E. coli strains (Pyo18 with deleted hlya-Pyo18ΔhlyA- and Pyo14) were compared. Following 4 h of incubation, Pyo18 decreased epithelial cell numbers to 54% (p < 0.001), and induced death of all stromal cells (p < 0.0001), whereas Pyo18ΔhlyA and Pyo14 had no effect on cell numbers. Compared to Pyo18ΔhlyA and Pyo14, respectively, Pyo18 induced a lower transcription level of IL-1β (0.99 vs 152.0 vs 50.9 fold increase, p < 0.001), TNFα (3.2 vs 49.9 vs 12.9 fold increase, p < 0.05) and IL-10 (0.4 vs 3.6 vs 2.6 fold increase, p < 0.001) in stromal cells, after 1 h of incubation. This may be seen as an attempt of hemolytic E. coli to delay the activation of the immune response. In conclusion, endometrial epithelial and stromal cell damage induced by HlyA is a potential relevant step of E. coli virulence in the pathogenesis of pyometra.
- Stromal Cell
- Systemic Inflammatory Response Syndrome
- Endometrial Tissue
- Endometrial Cell
- Endometrial Stromal Cell
Pyometra is a common diestrous disease of bitches. Escherichia coli is isolated from the uterus of up to 90% of bitches with pyometra [1, 2] and its presence is associated with severe systemic signs and a potentially life-threatening condition. The systemic inflammatory response syndrome (SIRS) is detected in more than 50% of bitches with E. coli pyometra . These E. coli isolates derive from the host’s fecal and perineal flora , being mainly assigned to phylogenetic group B2, and characterized by a high number of uropathogenic E. coli (UPEC) virulence factor (VF) genes and pathogenicity-associated islands markers . The α-hemolysin (hlyA) VF gene was detected in 35–52% of E. coli pyometra cases [2, 5, 6]. Although this prevalence leads to the suggestion that α-hemolysin (HlyA) contributes to the virulence of E. coli strains, the role of this toxin in the pathogenesis of canine pyometra is unknown. HlyA is a RTX pore-forming exotoxin. At high concentrations, HlyA is able to lyse erythrocytes and nucleated host cells. At sublytic concentrations, HlyA can disrupt the immune signaling and cytoskeletal components .
Toll-Like-Receptor (TLR)-mediated immune surveillance is an important component of the defence mechanisms within the canine uterus . Up-regulation of TLR2 and four transcription [9, 10] and expression  was observed in E. coli pyometra endometrium. Uterine inflammatory response towards E. coli is associated with an endometrial up-regulation of genes related with chemokines, cytokines, inflammatory cell extravasation, anti-bacterial action, proteases and innate immune response [9, 11]. In accordance, pyometra is characterized by endometrial tissue damage, infiltration by inflammatory cells, accumulation of pus, and increased expression of inflammatory mediators such as interleukins IL-1β, IL-6 and IL-8 [9, 11]. Different expression patterns of cytokines were observed in bitches with or without pyometra-associated SIRS .
Most of the studies on canine pyometra were carried out at the time of diagnosis, a late stage of the disease, and did not take into account the virulence background of E. coli. The characterization of the endometrial cell cytokine response to E. coli may lead to a relevant insight into the pathogenesis of pyometra. Additionally, the characterization of the role of specific E. coli VF genes in the modulation of the endometrial immune response and on the pathogenicity of the bacterium may prove rewarding in the development of novel diagnostic and therapeutic approaches to the disease. In this regard, HlyA becomes a promising candidate.
This study includes two experiments. Experiment 1 was designed to evaluate and compare the clinical and uterine histological and immune response gene transcription of hemolytic and non-hemolytic E. coli pyometra bitches. Prompted by results of Experiment 1, Experiment 2 was designed to evaluate the effects of E. coli HlyA on the modulation of the inflammatory response of canine endometrial epithelial and stromal cells.
Healthy diestrous (n = 10) and pyometra (n = 18) bitches presented to the Hospital of the Faculty of Veterinary Medicine of the University of Lisbon were selected for this experiment. The average age of bitches was 8 years (range 3–13 years). Healthy bitches were submitted to elective ovariohysterectomy (OHE) for contraceptive purposes, as requested by owners. In pyometra bitches the diagnosis was based on case history, clinical signs and the ultrasonographic finding of an enlarged, fluid-filled uterus. A blood sample for hematologic (Hemogram) and biochemical (urea, creatinine, phosphatase alkaline, alanine-aminotransferase) analysis was collected prior to OHE. After removing the uterus, an intra-uterine swab was processed for bacteriological analysis. Selection of the 18 pyometra bitches was based on the isolation of hemolytic (n = 8) and non-hemolitic (n = 10) E. coli strains. Uterine samples were collected as described previously  and either fixed for 24 h in 4% neutral phosphate buffered formalin (for IHC, described in sub-section “Immunohistochemistry”) or immersed for 24 h in RNAlater (Qiagen, GmbH, Hilden, Germany) and then stored at −80 °C (for RT-PCR and qRT-PCR, described in sub-section “RNA extraction, cDNA synthesis, RT-PCR and qRT-PCR”).
Immunohistochemistry (IHC) was used to identify T lymphocytes (Rabbit polyclonal anti-human CD3, diluted 1:200; Dako, Glostrup, Denmark), B lymphocytes (mouse monoclonal anti-human anti-CD79 αcy, clone HM57, diluted 1:150; Dako) and granulocytes and macrophages (mouse anti-human MCA874G, clone MAC387, diluted 1:400, Dako) in uterine samples. Except for anti-CD3 antibody, all protocol steps were carried out using the Novolink Polymer Detection System (Novocastra, Leica Biosystems, Newcastle, UK), according to the manufacturer’s instructions. The antigen retrieval step was performed by microwave treatment (3 × 5 min) in Tris–EDTA buffer (pH 9.0). After endogenous peroxidase blocking and treatment with protein block solution (Protein Block Solution-Kit NovoLink™), sections were incubated 1 h at room temperature with the respective primary antibodies. CD3 immunostaining was carried out as previously described  with minor modifications. Briefly, endogenous peroxidase was quenched by incubating the slides in 3% hydrogen peroxide in water for 30 min followed by antigen retrieval in Tris–EDTA buffer (pH 9.0), as described above. Blocking was performed with blocking solution (PBS + 0.1% Tween + 5% goat serum + 2.5% BSA), for 1 h at room temperature followed by incubation with the primary antibody for 2 h. The peroxidase conjugated goat anti-rabbit IgG polyclonal antibody (diluted 1:100, Dako) was used as secondary antibody. For all antibodies, the staining was developed by incubating the slides with the substrate diaminobenzidine (DAB kit, Zytomed Systems, Berlin, Germany). Staining without the primary antibody and staining with the isotype-matched irrelevant monoclonal antibody were used as negative controls. Human tonsil and dog spleen sections were used as positive control tissues.
Uteri were obtained from six healthy bitches (1–3 years old) submitted to ovariohysterectomy (OHE) for contraceptive purposes at the Teaching Hospital of the Faculty of Veterinary Medicine of the University of Lisbon. Only uteri from the first half of diestrus (as determined by vaginal cytology, blood progesterone concentrations and histology), without histological evidence of cystic endometrial hyperplasia (CEH) and a negative bacteriological result were allocated to the study. Three uteri were used for the evaluation of bacterial adhesion and internalization, and the remaining three uteri were used for the evaluation of cytokine transcription, IHC, and the number and morphology of cells following stimulation. All these evaluations considered two technical replicates per uteri for each stimulus (negative un-stimulated control (NUC), LPS, Pyo14, Pyo18, Pyo18ΔhlyA).
Endometrial epithelial and stromal cell populations were isolated as described by Bläuer et al.  and Stadler et al.  with some modifications. Briefly, small endometrial strips were digested with 1 mg/mL of collagenase A (Roche Diagnostics GmbH, Mannheim, Germany) and 5 μg/mL of DNAse I (Sigma-Aldrich, St. Louis, USA) in RPMI-1640 (GIBCO®, Invitrogen Corporation, New York, USA) supplemented with gentamicin (50 μg/mL, Sigma-Aldrich) and anfotericin B (2.5 μg/mL, Sigma-Aldrich). After 60-90 min of incubation with gently shaking at 37 °C, the cell suspension was filtered through a sterile mesh to remove undigested tissue. The suspension was centrifuged at 200 × g for 5 min and re-suspended in 10 mL of culture medium (RPMI-1640 supplemented with 10% heat inactivated fetal bovine serum, 50 μg/mL of gentamicin and 2.5 μg/mL de anfotericin B). Epithelial organoids and stromal fibroblasts were separated by differential centrifugation and filtered through a sterile 40 μm pore size filter. Stromal suspension (bottom of the filter) was pelleted by centrifugation (250 × g), freed of red blood cells and suspended in 1 mL of culture medium. Further purification of stromal cells was done by gently pipetting the fraction onto 9 mL of culture medium and letting them to sediment by gravity for 30 min.
Organoid suspension was obtained after back washing the filter with culture medium. After 2 steps of sedimentation by gravity for 5 min, epithelial cells suspension was centrifuged (200 × g, 5 min) and the pellet was suspended in 3 mL of trypsin 0.25%—EDTA (GIBCO®) with 5 μg/mL of DNAse I. After 2–5 min of incubation, organoids were unbundled with a 26G needle and 2 mL syringe in order to get a single-cell suspension. The viability of stromal and epithelial cells, as assessed by Trypan Blue exclusion dye staining before culture, was 80–95%.
Stromal and epithelial cells were seeded in 24-well plates at a density of 1 × 105 cells/mL and 2 × 105 cells/well/mL, respectively, and cultured in 1 mL of culture medium supplemented with Insulin-Transferrin-Selenium (1×) (ITS liquid media supplement, 100×, Sigma-Aldrich). For stromal cells, medium was changed 12 h after plating to allow selective attachment. For epithelial cells, medium was removed after 24 h to discard nonattached cells. Culture media was then changed every 48 h until cells reached 80% confluence. All cultures were incubated at 37 °C, 5% CO2 in air, in a humidified incubator. Purity of stromal and epithelial cells, as determined by their morphology after staining with Giemsa (Accustain®, Sigma-Aldrich) in glass coverslips, was respectively 95 and 85%.
E. coli strains
The two E. coli strains used in this study, Pyo14 and Pyo18, were previously isolated from the uterus of pyometra bitches and characterized . Although both strains were from phylogenetic group B2, only Pyo18 carries hlyA gene and have a hemolytic phenotype in sheep blood agar . The isogenic hlyA deletion mutant was generated in E. coli Pyo18 strain using the Lambda-Red recombinase system as described by Datsenko and Wanner . Briefly, the chloramphenicol cassette was amplified from a PKD3 plasmid using primers with 40 nucleotide extension homologous to the hlyA gene (hlyA1: 5′-cagatttcaatttttcattaacaggttaagaggtaattaagtgtaggctggagctgcttc-3′; hlyA2: 5′-cagcccagtaagattgctattatttaaattaataaaatgggaattagccatggtcc-3′). The following PCR conditions were used: initial denaturation of 5 min at 95 °C, 35 cycles of 30 s at 94 °C, 30 s at 50 °C, 2 min at 72 °C and a final extension of 5 min at 72 °C. A purified PCR product was used to replace the chromosomal hlyA gene in Pyo18 strain using the helper plasmid pKD46 expressing the lambda recombinase. For this purpose, Pyo18 strain was previously transformed with pKD46 plasmid, which after induction with 0.01 M l-arabinose promotes homologous recombination between the PCR product and the chromosomal target gene. The PCR product was electroporated into Pyo18/PKD46 and the chloramphenicol resistance gene was eliminated using the helper plasmid pCP20. Elimination of pCP20 plasmid was accomplished by bacterial growth at 43 °C. The hlyA gene deletion was confirmed by PCR analysis (∆hlyA1: 5′-ccattagaggttcttgggc-3′; ∆hlyA2: 5′-ggaataaaccaggtaaagtc-3′) and DNA sequencing.
Before stimulation experiments, bacteria were plated onto Columbia agar medium and incubated at 37 °C overnight. For each bacterium, one bacterial colony was inoculated into liquid LB medium at 37 °C for 48 h without agitation, washed three times by centrifugation (5000 × g for 5 min) and re-suspended in sterile PBS. The bacterial suspension was diluted at a final working concentration equivalent to 108 CFU/mL. Similar bacterial growth curves were obtained for the three E. coli strains (two replicates per strain).
E. coli cell stimulation
Cell culture conditions were maintained during stimulation. After reaching 80% of confluence, cells were pre-incubated in medium without antibiotic (RPMI-1640 supplemented with 10% of FBS) during 24 h before bacteria inoculation. Endometrial epithelial and stromal cells were incubated with the E. coli isolates [multiplicity of infection (MOI) of 6–8], LPS (1 μg/mL E. coli O55:B5, Sigma-Aldrich) or medium alone for 1 and 4 h at 37 °C (adhesion step). Cells were washed with PBS to remove non-adherent E. coli and further incubated for 1.5 h with fresh medium containing 50 μg/mL gentamicin to eliminate extracellular bacteria (internalization step). The medium was then removed and the cells washed three times with PBS. Cells of two duplicate wells were lysed with 350 µL of RLT/b-ME buffer (Rneasy mini kit, Qiagen GmbH) and frozen at −80 °C until RNA extraction. In parallel plates, after internalization, cells were washed and cell morphology was evaluated. Cells in each well were then trypsinized and the number of viable cells was counted in a Neubauer chamber, using a contrast phase equipped microscope. Results of cell counts for each type of stimulation are presented as the mean percentage relative to the respective unstimulated control.
Adherence and internalization levels were evaluated as described by Letourneau et al. . The percentage of adhered bacteria was calculated by dividing the number of cfu of adhered bacteria by the addition of the number of cfu of adhered and non-adhered bacteria. The percentage of internalization was calculated by dividing the number of cfu of internalized cells by the number of cfu of adhered cells.
RNA extraction, cDNA synthesis, RT-PCR and qRT-PCR
Total RNA was isolated from pyometra endometrial tissue and from cell cultures using the Rneasy mini kit (Qiagen GmbH) and DNA digestion was performed with the RNase-free DNase set (Qiagen GmbH). RNA concentration and purity was determined in a NanoDrop 2000/2000c spectrophotometer (Thermo Fisher Scientific, Walthman, USA). Single-stranded complementary DNA (cDNA) synthesis was performed by reverse transcription of 500 ng (endometrial tissue) or 400 ng (cell cultures) of total RNA using the SuperScript III First-Strand synthesis SuperMix for qRT-PCR (Invitrogen, Austin, TX, USA), according to the manufacturers’ protocols.
Primer sequences used for RT-PCR
Annealing temperature (°C)
Amplicon size (bp)
Primers sequences used for quantitative real time PCR (qRT-PCR)
Amplicon size (bp)
Real-time PCR was performed in duplicate wells on StepOnePlus™ (Applied Biosystems), using the universal temperature cycles as suggested by the manufacturer. Melting curves were acquired to ensure that a single product was amplified in the reaction. All PCR reactions were carried out in 96-well optical reaction plates (Applied Biosystems, Warrington, UK) with 1× Power SYBR® Green PCR Master Mix (Applied Biosystems), 1 ng (endometrial tissue) or 2 ng (cell cultures) of cDNA, 80 nM of each primer in a total reaction volume of 12.5 μL. After analyzing the melting curves, the PCR products were run through a 2.5% gel agarose to confirm the expected product size. The identity of PCR products was initially confirmed by DNA sequencing. The data of relative mRNA quantification was analyzed with the real-time PCR miner algorithm  and with 2−ΔΔCT method (for IL-10, INFβ, IL-1α) . Results of qRT-PCR are expressed as the fold increase relative to the unstimulated control.
After the stimulation assay, epithelial and stromal cells were washed three times in sterile PBS, fixed for 10 min in 4% paraformaldehyde (Sigma), washed two times in PBS- 0.1% Triton for 5 min and incubated with the blocking solution (PBS + 2.5% BSA) for 1 h at room temperature. Cells were incubated overnight at 4 °C with rabbit polyclonal anti-IRF3 antibody (1:50, ab25950, Abcam®, Cambridge, UK) and rabbit polyclonal anti-NFkB p65 antibody (1:100, sc-109-G, Santa Cruz biotechnology®, Texas, USA). After washing twice for 10 min in PBS, the secondary antibody (1:300, Alex Fluor® 594 Goat Anti-Rabbit IgG (H + L), Life Technologies, Carlsbad, USA) was added and cells incubated for 1 h at room temperature. After washing for 10 min in PBS, the glass coverslips were mounted in Vectashield® mounting medium with DAPI (H-1200, Vector Laboratories Inc., Burlingame, CA, USA). Slides were observed using a fluorescence microscope (Leica DM5000B) and the images obtained in Adobe Photoshop CS5.
Data were analyzed through a statistical software package (Statistica 5.0, StatSoft Inc., Tulsa, OK, USA, 1995). Categorical data was analyzed by Fisher’s exact test. Hemogram, blood biochemical, cell number and cytokine transcription data were log-transformed (log x + 1) to normalize distribution. Hemogram, blood biochemical and endometrial cytokine transcription data (Experiment 1) and cell number (Experiment 2) were analyzed by ANOVA. In stromal cells Pyo18 induced cell death after 4 h of incubation, so transcription data was not available. Therefore, two different datasets were considered in Experiment 2. One dataset excluded data from stimulation with Pyo18. This dataset was analyzed using the MANOVA factorial procedures with three fixed effects: type of cells (n = 2; stromal cells, epithelial cells); type of stimulus (n = 3; LPS, Pyo14, Pyo18ΔhlyA); time of incubation (1, 4 h) and their interactions. Additionally, a second dataset excluded data from the 4-h endpoint. This dataset was analyzed by a MANOVA factorial procedure with two fixed effects: type of cells (n = 2; stromal cells, epithelial cells) and type of stimulus (n = 4; LPS, Pyo14, Pyo18ΔhlyA, Pyo18). Significant effects were further analyzed and means compared using the Scheffé test. Significance was determined at the 5% confidence level (p < 0.05).
Hemolytic E. coli induces a more extensive uterine damage and inflammatory cell infiltration than non-hemolytic E. coli (Experiment 1)
Hematological and biochemical parameters in bitches with hemolytic E. coli pyometra ( n = 8) and with non-hemolytic E. coli pyometra ( n = 10)
Hemolytic E. coli pyometra (Mean ± SEM)
Non-hemolytic E. coli pyometra (Mean ± SEM)
White blood cell count (109 l−1)
45.1 ± 5.5
33.8 ± 9.1
Eritrocytes (×1012 l−1)
5.5 ± 0.1
6.1 ± 0.4
Platelet count (109 l−1)
216.1 ± 66.3
247.4 ± 48.7
Hemoglobin (g l−1)
126.0 ± 2.8
137.4 ± 7.8
37.3 ± 0.9
40.2 ± 2.4
Mean corpuscular volume (fL)
67.6 ± 0.9
66.1 ± 1.0
Mean corpuscular hemoglobin (pg)
22.9 ± 0.4
22.6 ± 0.2
Mean corpuscular hemoglobin concentration (g/dl)
33.9 ± 0.6
34.2 ± 0.4
Neutrophils (109 l−1)
34.8 ± 4.2
25.8 ± 7.3
Band Neutrophils (109 l−1)
0.9 ± 0.7
2.2 ± 1.1
Lymphocytes (109 l−1)
3.4 ± 0.5
2.3 ± 0.4
Monocytes (109 l−1)
5.65 ± 1.8
3.2 ± 0.9
Eosinophils (109 l−1)
0.2 ± 0.1
0.3 ± 0.2
Basophils (109 l−1)
0.3 ± 0.3
Alanine-aminotransferase (U L−1 at 25 °C)
22.5 ± 3.9
38.1 ± 12.5
Alkaline phosphatase (U L−1)
309.2 ± 146.5
488.1 ± 311.8
Creatinine (μmol L−1)
114.9 ± 17.7
97.2 ± 8.8
Urea (mmol L−1)
5.1 ± 1.6
6.4 ± 2.9
An extensive stromal infiltration of myeloid cells (granulocytes/macrophages) (Figures 1D and E) and B lymphocytes was observed, being B lymphocytes more predominant in the apical than in the basal layer (Figures 1G and H). Stromal infiltration of T lymphocytes was moderate and more predominant in the apical layer (Figures 1J and K). Most (84%) of the pyometra samples also showed an extensive infiltration of myeloid cells in endometrial glands, whereas the glandular infiltration of lymphocytes was absent or weak. The presence of inflammatory cells in normal diestrous uteri was scarce (detected in 30% of the samples) and consisted of few granulocytes/macrophages (apical and basal layer) and T cells (apical layer) in the stromal compartment.
Hemolytic and non-hemolytic E. coli pyometra endometria have different transcription levels of genes coding pro-inflammatory cytokines (Experiment 1)
Overall, transcription levels of pro-inflammatory cytokines IL-1β, IL-6 and IL-8, and of anti-inflammatory cytokines IL-10 and TGFβ were significantly higher (p < 0.0001) in pyometra endometria than in healthy diestrous endometria. However, hemolytic E. coli pyometra endometria had higher transcription levels of IL-1β and IL-8 and lower transcription levels of IL-6 than non-hemolytic E. coli endometria (p < 0.01) (Figure 2B).
Cytotoxicity due to α-hemolysin is mainly targeted to stromal endometrial cells (Experiment 2)
In epithelial cells, bacterial adhesion after 1 and 4 h of incubation was similar (3-5%) for both E. coli strains. This low bacterial adhesion was also observed in stromal cells incubated with Pyo14. In both epithelial and stromal cells incubated with Pyo14, bacterial internalization was absent (at 1 h) or very low (at 4 h; 0.1 – 0.2%). Incubation of stromal cells with Pyo18 for 4 h resulted in the death of most cells, therefore bacterial adhesion and internalization could not be evaluated.
Effect of bacterial or LPS incubation on endometrial epithelial and stromal cell numbers
E. coli strain
1 h (%)
4 h (%)
1 h (%)
4 h (%)
102.7 ± 5.9
98.9 ± 2.5*
107.0 ± 1.4
96.3 ± 6.3*
102.9 ± 4.8a
105.9 ± 3.3c
54.3 ± 3.4d,**
104.5 ± 3.0
97.4 ± 3.2*
110.1 ± 2.3
97.9 ± 4.4*
100.7 ± 8.4
96.0 ± 4.6*
115.8 ± 2.2
99.9 ± 6.4*
In contrast, incubation with the isogenic mutant Pyo18ΔhlyA did not induce a cytotoxic effect on stromal cells (Figures 4C, F, I, L). In fact, following incubation with this mutant strain, cell morphology and numbers were similar to those observed after incubation with Pyo14 (Table 4; Figures 3 and 4).
Hemolytic and non-hemolytic E. coli strains induce differential immune response in endometrial epithelial and stromal cells
Mean cycle threshold (Ct) values of cytokine gene transcription in unstimulated endometrial epithelial and stromal cells following 1 and 4 h of incubation
In stromal cells, after 1 h of incubation and compared to the isogenic mutant Pyo18ΔhlyA and to Pyo14, respectively, Pyo18 induced a lower transcription level of IL-1β (0.99 vs 152.0 vs 50.9 fold increase, p < 0.001) (Figure 6A), TNFα (3.2 vs 49.9 vs 12.9 fold increases, p < 0.05) (Figure 6D) and IL-10 (0.4 vs 3.6 vs 2.6 fold increases, p < 0.001) (Figure 7C). Stimulation with Pyo14, induced higher IL-1α and IL-8 transcription levels after 4 h than 1 h of incubation (1 h—2.2 vs 4 h—7.9 fold increase, p < 0.05 and 1 h—19.8 vs 4 h—184.1 fold increase, p < 0.01, respectively) (Figures 6B and 7A). Stimulation with Pyo18ΔhlyA also induced higher IL-1α and IL-8 transcription levels after 4 h than 1 h of incubation (1 h—4.5 vs. 4 h—9.9 fold increase, p < 0.05 and 1 h—4.5 vs. 4 h—168.7 fold increase, p < 0.01, respectively) (Figures 6B and 7A). Transcription of INFβ was not detected in stromal cells.
The bacterial stimuli induced different gene transcription levels in endometrial epithelial and stromal cells. Pyo14 and Pyo18ΔhlyA induced higher transcription levels of CXCL10 after 4 h of incubation in epithelial than in stromal cells (Pyo14—146.5 vs 58.9 fold increase, p < 0.05 and Pyo18Δhly—142.4 vs 44.9 fold increase, p < 0.01, respectively) (Figure 7B). Also, Pyo14 and Pyo18ΔhlyA induced higher transcription levels of IL-8 in stromal than epithelial cells after 4 h of incubation (Pyo14—184.7 vs 6.7 fold increase; Pyo18ΔhlyA—168.7 vs 19.8 fold increase, respectively, p < 0.02) (Figure 7A).
The results of this study indicate that HlyA contributes to the virulence of E. coli, by inducing tissue damage and a compromised early uterine immune response. On one side, the cytotoxic activity towards epithelial and stromal cells may facilitate progression of bacteria into the uterine tissue. On the other side, the modulation of the uterine innate immune response may induce a delayed neutrophil influx to the uterus. Additionally, the known cytotoxic effect of HlyA in granulocytes and lymphocytes, as well as the enhancement of access to cell nutrients and iron storages , might also enhance hemolytic E. coli survival within the uterine tissue.
The relationship between the isolation of β-hemolytic E. coli and clinical signs and histopathological findings was not reported. UPEC strains that express HlyA were shown to cause more extensive tissue damage within the urinary tract, which was correlated with more severe clinical outcomes . In this study, clinical signs, hematological and blood biochemical results were similar in hemolytic and non-hemolytic E. coli pyometra bitches. However, β-hemolytic E. coli was associated with high endometrial damage and metritis.
All E. coli pyometra uteri evidenced an infiltrate of inflammatory cells, including neutrophils, lymphocytes and macrophages. This may contribute for the up-regulation of TLR2 and TLR4 transcription [9, 10] and expression  and the consequent up-regulation of gene transcription of several pro- and anti-inflammatory cytokines, including IL-1β, IL-6, IL-8, IL-10 and TGFβ. However, hemolytic E. coli endometria had higher transcription levels of IL-1β and IL-8 and lower transcription levels of IL-6 than non-hemolytic E. coli endometria. IL-6 stimulates leukocyte activation and myeloid progenitor cell proliferation, regulates the synthesis of acute phase proteins and is a powerful pyrogen . In pyometra cases, C-reactive protein, serum amyloid A and haptoglobin concentrations can be used to monitor early post-OHE complications and ongoing inflammation . C-reactive protein was also associated with SIRS . However, in the above studies the hemolytic nature of E. coli was not evaluated. High serum levels of IL-8 were observed in pyometra bitches, especially in those that developed SIRS, suggesting IL-8 as a useful early biomarker of uterine infection and, possibly, of sepsis . IL-8 is also a potent neutrophil chemotactic molecule. Neutrophil recruitment to the site of infection was shown to be critical for bacterial clearance in the uterus, although this may also lead to tissue damage .
Endometrial epithelial cells showed a high sensitivity to the cytotoxic effect of HlyA. This may lead to the suggestion that β-hemolytic E. coli induces an earlier damage to the epithelial layer than non-hemolytic E. coli strains. During uterine contamination, epithelial cell apoptosis  and/or degradation of proteins involved in cell–cell and cell–matrix interactions  due to the sublytic concentrations of HlyA may be the reason for the disruption of the epithelial layer. Additionally, the cytotoxic effect of HlyA may be associated with the observed destruction of epithelial glandular cells of pyometra uteri. Endometrial stromal cells showed a higher sensitivity to the cytotoxic effect of HlyA than epithelial cells. A higher sensitivity of bovine stromal cells to pyolysin-mediated cytolysis of Trueperella pyogenes was recently reported . This sensitivity of both epithelial and stromal cells to the cytotoxic effect of HlyA may explain the observed significant association between β-hemolytic E. coli isolation and metritis.
Adhesion to epithelial and stromal cells was similar following bacterial incubation, irrespective of E. coli strain. The observed level of E. coli adhesion is similar to that reported for bovine endometrial cells . Pyo14 and Pyo18 have genes for Type 1 fimbrae and S + F1C fimbriae, whereas the papG gene that codes for P fimbriated was only detected in Pyo18 . Although adherence of E. coli to canine endometrial cells was shown to be mediated by Type 1 fimbriae , the targeted deletion of specific adhesin genes in a canine pyometra strain was compensated by the presence of other adhesins, which indicates functional redundancy among adhesins . Bacterial internalization was absent or very low. This indicates that cell invasion is not necessary for inducing the cytotoxic effect of HlyA. Poor internalization of hemolytic E. coli was reported in HeLa and HTB-4 cell co-culture systems . Low internalization was also observed in bovine endometrial epithelial and stromal cells following incubation with non-hemolytic E. coli during 4 h .
Canine endometrial epithelial and stromal cells express TLR4  and TLR4 binding triggers a signaling cascade that culminates in the activation of NFkB and IRF3. The detection of NFkB and IRF3 in the nucleus, together with the observed cytokine transcription results, indicates that the TLR signaling pathway was activated.
Differential transcription of cytokine genes by endometrial stromal but not epithelial cells was observed following stimulation with hemolytic and non-hemolytic E. coli strains. In stromal cells, transcription levels of TNFα, IL-1β and IL-10 were lower following stimulation with Pyo18 than following stimulation with Pyo18ΔhlyA and Pyo14. The decrease in TNFα is probably responsible for the observed low levels of IL-1β. König and König  reported a down-regulation of TNFα expression in human inflammatory cells stimulated with hemolytic E. coli. Recently, in a bacteremia mouse model, HlyA inhibited IL-1β secretion . A functional impact of early IL-10 in UPEC UTI mice model were suggested as IL10 − /− mice had more UPEC and more severe bacteriuria . Overall, the observed down-regulation in TNFα, IL-1β and IL-10 transcription in stromal cells may be seen as an attempt of hemolytic E. coli to delay the activation of the immune response.
Differential cytokine gene transcription was also observed in endometrial epithelial and stromal cells. The transcription of CXCL10 was significantly higher in endometrial epithelial than stromal cells. CXCL10 displays antimicrobial activity against E. coli , and likewise defensins, may act as the first line of defense in epithelia . A high transcription level of CXCL10 was observed in canine endometrial stromal cells after stimulation with E. coli and LPS . In stromal cells, CXCL10 may be involved in the recruitment and potentiation of T-helper 1 response . However, γδ+/CD8− T lymphocytes were the only significant lymphocyte population detected in pyometra uteri .
Stromal cells had a higher transcription level of IL-8 than epithelial cells after incubation for 4 h with Pyo14 and Pyo18ΔhlyA. A high transcription level of IL-8 in canine stromal cells after E. coli and LPS stimulation was reported by Karlsson et al. . This may indicate that, during uterine E. coli infection and after disruption of the epithelial layer of the endometrium, stromal cells are an important early source of IL-8, a major cytokine involved in the activation and recruitment of neutrophils.
Transcription of INFβ was not detected in endometria (Experiment 1), but was detected in endometrial epithelial cells (Experiment 2) after incubation with Pyo18, Pyo18ΔhlyA, Pyo14 and LPS. Transcription of IFNβ by endometrial epithelial cells after 1 h of bacterial stimuli is likely triggered by LPS and is probably responsible for the up-regulation of CXCL10 transcription after 4 h of stimulation. The lack of transcription of IFNγ in normal diestrus and pyometra uteri is in accordance with the results of endometrial transcription and serum levels of IFNγ found in pyometra cases [3, 9].
In conclusion, β-hemolytic E. coli infection was associated with the occurrence of metritis and with higher uterine tissue damage than non-hemolytic E. coli infection. β-hemolytic E. coli pyometra endometria had higher gene transcription of IL-1β and IL-8 and lower gene transcription of IL-6 than non-hemolytic pyometra endometria. The cytotoxic effect of HlyA was more severe in endometrial stromal than in epithelial cells. Endometrial stromal cell damage by HlyA is a potential relevant step of E. coli virulence in the pathogenesis of pyometra. Additionally, the HlyA-induced decrease of cytokine transcription by stromal cells, potentially leading to a decrease of cytokine expression, which may occur during the early stages of infection, may allow bacteria to establish a niche prior to the activation of an adequate innate immune response.
The authors declare that they have no competing interests.
Conceived and designed the experiments: ES, LM. Performed the experiments: SH, ES, MFS, SC, PD, LM. Analyzed the data: SH, ES, MFS, SC, LC, LM. Wrote the paper: SH, ES, LC, LM. All authors read and approved the final manuscript.
The authors thank Professor Conceição Peleteiro and Dr Hugo Pissarra of the Pathology Laboratory of CIISA, for their help in the histological analysis.
This work was supported by Grants PTDC/CVT/66587/2006 and UID/CVT/00276/2013 from Foundation for Science and Technology (FCT). Sofia Henriques is a Ph.D. fellow from FCT (SFRH/BD/65044/2009).
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
- Dhaliwal GK, Wray C, Noakes DE (1998) Uterine bacterial flora and uterine lesions in bitches with cystic endometrial hyperplasia (pyometra). Vet Rec 143:659–661PubMedGoogle Scholar
- Mateus L, Henrique S, Merino C, Pomba C, Lopes da Costa L, Silva E (2013) Virulence genotypes of Escherichia coli canine isolates from pyometra, cystitis and fecal origin. Vet Microbiol 166:590–594View ArticlePubMedGoogle Scholar
- Karlsson I, Hagman R, Johannisson A, Wang L, Karlstam E, Wernersson S (2012) Cytokines as immunological markers for systemic inflammation in dogs with pyometra. Reprod Domest Anim 47(Suppl 6):337–341View ArticlePubMedGoogle Scholar
- Hagman R, Kühn I (2002) Escherichia coli strains isolated from the uterus and urinary bladder of bitches suffering from pyometra: comparison by restriction enzyme digestion and pulsed-field gel electrophoresis. Vet Microbiol 84:143–153View ArticlePubMedGoogle Scholar
- Chen YMM, Wright PJ, Lee C-S, Browning GF (2003) Uropathogenic virulence factors in isolates of Escherichia coli from clinical cases of canine pyometra and feces of healthy bitches. Vet Microbiol 94:57–69View ArticlePubMedGoogle Scholar
- Siqueira A, Ribeiro M, Leite D, Tiba M, Moura C, Lopes M, Prestes N, Salerno T, Silva A (2009) Virulence factors in Escherichia coli strains isolated from urinary tract infection and pyometra cases and from feces of healthy dogs. Res Vet Sci 86:206–210View ArticlePubMedGoogle Scholar
- Wiles TJ, Mulvey MA (2013) The RTX pore-forming toxin alpha-hemolysin of uropathogenic Escherichia coli: progress and perspectives. Future Microbiol 8:73–84View ArticlePubMedPubMed CentralGoogle Scholar
- Silva E, Henriques S, Brito S, Ferreira-Dias G, Lopes-da-Costa L, Mateus L (2012) Oestrous cycle-related changes in production of Toll-like receptors and prostaglandins in the canine endometrium. J Reprod Immunol 96:45–57View ArticlePubMedGoogle Scholar
- Hagman R, Rönnberg E, Pejler G (2009) Canine uterine bacterial infection induces upregulation of proteolysis-related genes and downregulation of homeobox and zinc finger factors. PLoS One 4:e8039View ArticlePubMedPubMed CentralGoogle Scholar
- Silva E, Leitão S, Henriques S, Kowalewski MP, Hoffmann B, Ferreira-Dias G, Lopes da Costa L, Mateus L (2010) Gene transcription of TLR2, TLR4, LPS ligands and prostaglandin synthesis enzymes are up-regulated in canine uteri with cystic endometrial hyperplasia-pyometra complex. J Reprod Immunol 84:66–74View ArticlePubMedGoogle Scholar
- Voorward F, Marchi F, Villacis R, Alves C, Toniollo G, Amorim R, Drigo S, Rogatto S (2015) Molecular expresson profile reveals potential biomarkers and therapeutic targets in canine endometrial lesions. PLoS One 10:e0133894View ArticleGoogle Scholar
- Bläuer M, Heinonen PK, Martikainen PM, Tomás E, Ylikomi T (2005) A novel organotypic culture model for normal human endometrium: regulation of epithelial cell proliferation by estradiol and medroxyprogesterone acetate. Hum Reprod 20:864–871View ArticlePubMedGoogle Scholar
- Stadler K, Handler J, Schoenkypl S, Walter I (2009) A three-dimensional culture model of canine uterine glands. In Vitro Cell Dev Biol Anim 45:35–43View ArticlePubMedGoogle Scholar
- Henriques S, Silva E, Lemsaddek A, Lopes-da-Costa L, Mateus L (2014) Genotypic and phenotypic comparison of Escherichia coli from uterine infections with different outcomes: clinical metritis in the cow and pyometra in the bitch. Vet Microbiol 170:109–116View ArticlePubMedGoogle Scholar
- Datsenko KA, Wanner BL (2000) One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc Natl Acad Sci U S A 97:6640–6645View ArticlePubMedPubMed CentralGoogle Scholar
- Letourneau J, Levesque C, Berthiaume F, Jacques M, Mourez M (2011) In vitro assay of bacterial adhesion onto mammalian epithelial cells. J Vis Exp 51:2783Google Scholar
- Zhao S, Fernald RD (2005) Comprehensive algorithm for quantitative real-time polymerase chain reaction. J Comput Biol 12:1047–1064View ArticlePubMedPubMed CentralGoogle Scholar
- Livak KJ, Schmittgen TD (2001) Analysis of relative gene expression data using real-time quantitative PCR and the 2− ΔΔCT method. Methods 25:402–408View ArticlePubMedGoogle Scholar
- Bien J, Sokolova O, Bozko P (2012) Role of uropathogenic Escherichia coli virulence factors in development of urinary tract infection and kidney damage. Int J Nephrol 2012:681473View ArticlePubMedPubMed CentralGoogle Scholar
- Smith YC, Rasmussen SB, Grande KK, Conran RM, O’Brien AD (2008) Hemolysin of uropathogenic Escherichia coli evokes extensive shedding of the uroepithelium and hemorrhage in bladder tissue within the first 24 h after intraurethral inoculation of mice. Infect Immunol 76:2978–2990View ArticleGoogle Scholar
- Scheller J, Chalaris A, Schmidt-Arras D, Rose-John S (2011) The pro and anti-inflammatory properties of the cytokine interleukin-6. Biochim Biophys Acta 1813:878–888View ArticlePubMedGoogle Scholar
- Dabrowski R, Kostro K, Lisiecka U, Szczubiał M, Krakowski L (2009) Usefulness of C-reactive protein, serum amyloid A component, and haptoglobin determination in bitches with pyometra for monitoring early post-ovariohysterectomy complications. Theriogenology 72:471–476View ArticlePubMedGoogle Scholar
- Fransson BA, Lagerstedt A-S, Bergstrom A, Hagman R, Park JS, Chew BP, Evans MA, Ragle CA (2007) C-reactive protein, tumor necrosis factor a, and interleukin-6 in dogs with pyometra and SIRS. J Vet Emerg Critical Care 17:373–381View ArticleGoogle Scholar
- Kruger P, Saffarzadeh M, Weber ANR, Rieber N, Radsak M, von Bernuth H, Benarafa C, Roos D, Skokowa J, Hartl D (2015) Neutrophils: between host defence, immune modulation, and tissue injury. PLoS Pathog 11:e1004651View ArticlePubMedPubMed CentralGoogle Scholar
- Smith YC, Grande KK, Rasmussen SB, O’Brien AD (2006) Novel three-dimensional organoid model for evaluation of the interaction of uropathogenic Escherichia coli with terminally differentiated human urothelial cells. Infect Immunol 74:750–757View ArticleGoogle Scholar
- Dhakal BK, Mulvey MA (2012) The UPEC pore forming toxin α-hemolysin triggers proteolysis of host proteins to disrupt cell adhesion, inflammatory and survival pathways. Cell Host Microbe 11:58–69View ArticlePubMedPubMed CentralGoogle Scholar
- Amos M, Healey G, Goldstone R, Mahan S, Duvel A, Schuberth H, Sandra O, Zieger P, Dieuzy-Labaye I, Smith D, Sheldon I (2014) Differential endometrial cell sensitivity to a cholesterol-dependent cytolysin links Trueperella pyogenes to uterine disease in cattle. Biol Reprod 90:54View ArticlePubMedGoogle Scholar
- Sheldon IM, Rycroft AN, Dogan B, Craven M, Bromfield JJ, Chandler A, Simpson KW (2010) Specific strains of Escherichia coli are pathogenic for the endometrium of cattle and cause pelvic inflammatory disease in cattle and mice. PLoS One 5:e9192View ArticlePubMedPubMed CentralGoogle Scholar
- Krekeler N, Marenda MS, Browning GF, Holden KM, Charles JA, Wright PJ (2012) Uropathogenic virulence factor FimH facilitates binding of uteropathogenic Escherichia coli to canine endometrium. Comp Immunol Microbiol Infect Dis 35:461–467View ArticlePubMedGoogle Scholar
- Krekeler N, Marenda MS, Browning GF, Holden KM, Charles JA, Wright PJ (2013) The role of Type 1, P and S fimbriae in binding of Escherichia coli to the canine endometrium. Vet Microbiol 164:399–404View ArticlePubMedGoogle Scholar
- Saldaña Z, De la Cruz MA, Carrillo-Casas EM, Durán L, Zhang Y, Hernández-Castro R, Puente JL, Daaka Y, Girón JA (2014) Production of the Escherichia coli common pilus by uropathogenic E. coli is associated with adherence to HeLa and HTB-4 cells and invasion of mouse bladder urothelium. PLoS One 9:e101200View ArticlePubMedPubMed CentralGoogle Scholar
- Konig B, Konig W (1993) Induction and suppression of cytokine release (TNF-α; IL-6; IL-1β) by E. coli pathogenicity factors (adhesins, alpha-hemolysin). Immunol 78:526–533Google Scholar
- Diabate M, Munro P, Garcia E, Jacquel A, Michel G, Obba S, Goncalves D, Luci C, Marchetti S, Demon D, Degos C, Bechah Y, Mege JL, Lamkanfi M, Auberger P, Gorvel JP, Stuart LM, Landraud L, Lemichez E, Boyer L (2015) Escherichia coli α-hemolysin counteracts the anti-virulence innate immune response triggered by the Rho GTPase activating toxin CNF1 during bacteremia. PLoS Pathog 11:e1004732View ArticlePubMedPubMed CentralGoogle Scholar
- Duell BL, Carey AJ, Tan CK, Cui X, Webb R, Totsika M, Schembri MA, Derrington P, Irving-Rodgers H, Brooks Andrew J, Cripps AW, Crowley M, Ulett GC (2012) Innate transcriptional networks activated in bladder in response to uropathogenic Escherichia coli drive diverse biological pathways and rapid synthesis of IL-10 for defense against bacterial urinary tract infection. J Immunol 188:781–792View ArticlePubMedGoogle Scholar
- Cole AM, Ganz T, Liese AM, Burdick MD, Liu L, Strieter RM (2001) Cutting edge: IFN-inducible ELR-CXC chemokines display defensin-like antimicrobial activity. J Immunol 167:623–627View ArticlePubMedGoogle Scholar
- Gallo RL, Murakami M, Ohtake T, Zaiou M (2002) Biology and clinical relevance of naturally occurring antimicrobial peptides. J Allergy Clin Immunol 110:823–831View ArticlePubMedGoogle Scholar
- Karlsson I, Hagman R, Guo Y, Humblot P, Wang L, Wernersson S (2015) Pathogenic Escherichia coli and lipopolysaccharide enhance the expression of IL-8, CXCL5, and CXCL10 in canine endometrial stromal cells. Theriogenology 84:34–42View ArticlePubMedGoogle Scholar
- Dufour JH, Michelle D, Liu MT, Leung JH, Lane TE, Luster AD (2002) IFN-gamma inducible protein 10 (IP-10; CXCL10)-deficient mice reveal a role for IP-10 in effector T cell generation and trafficking. J Immunol 168:3195–3204View ArticlePubMedGoogle Scholar
- Bartoskova A, Turanek-Knotigova P, Matiasovic J, Oreskovic Z, Vicenova M, Stepanova H, Ondrackova P, Vitasek R, Leva L, Moore PF, Faldyna M (2012) γδT lymphocytes are recruited into the inflamed uterus of bitches suffering from pyometra. Vet J 19:303–308View ArticleGoogle Scholar