Transport conditions from the slaughterhouse to the laboratory (transportation time approximately 2 h) turned out to be a crucial factor for cell viability. Tissue that was transported in growth medium at room temperature showed best survival of the epithelial cells.

Pure epithelial cervical cells could reproducibly be isolated only by outgrowth from tissue explants (Figure 1) as other previously described cell isolation methods did not lead to a pure and viable primary cell population in our hands.

Primary cells can be passaged and cryopreserved using standard protocols.

After passaging the cervical cells up to five times, the cultured cells were still viable and showed some epithelial characteristics (keratin staining). However, the epithelial morphology was completely lost in passage 5 (Figure 2). Already in passage one morphological alterations can be found in some parts of the monolayer (Figure 3).

Growth media were evaluated using a simple morphology score. The growth medium Ham's F-12 containing 10% FCS, antibiotics, antimycotics and antioxidants gained the highest score of 180 points concerning proliferation, the absence of vacuoles and epitheloid phenotype of the grown-out primary cells and cells in passage one (Table 1).

Conditioning of the growth medium and the addition of media supplements were investigated using cervical cells of passage one. Cells cultured in conditioned Ham's F-12 showed an increase in mitochondrial turnover of WST-1 of 12% after 48 h and 44% after 72 h relative to the unconditioned growth medium. Supplementation of EGF and porcine insulin to the conditioned Ham's F-12 led to a two-fold higher mitochondrial activity after 48 h and 72 h compared to the conditioned growth medium without any hormone substitution (Figure 4). Other additives or other additive combinations led to a less pronounced increase (EGF, EGF + hydrocortisone, EGF + hydrocortisone + insulin) or stagnation (insulin, hydrocortisone, insulin + hydrocortisone) of mitochondrial activity.

Primary cervical cells and cells of passage one were cultured on Millicell inserts either with access to growth medium from both basolateral and apical or in air-liquid interface with access to growth medium only from basolateral. We analyzed the respective setups histologically after three weeks considering in vivo-like morphology.

A differentiated phenotype was established when culturing primary cells in an air-liquid interface (without medium in the upper compartment) for three weeks. The cells showed a reproducible discontinuously stratified phenotype with up to six layers. When cultured conventionally with access to growth medium from both basolateral and apical the epithelial cells were mainly columnar and also showed partially multilayered growth. However, height of the multilayer was increased by growth in air liquid interface (conventional growth: up to 4 layers, air-liquid interface: up to six layers). After passaging, the epithelial cells grew as single-layer consisting of non-polarized cells. Passaged cells showed multilayered growth only in very few areas of the membranes (Figure 5).

Purity of the epithelial cell culture and differentiation was shown by immunohistochemistry using antibodies against keratin, as epithelial marker, and beta catenin, as part of adherens junctions. Cervical epithelial cells cultured on membranes in air-liquid interface for three weeks showed keratin-expression in the cytoplasm comparable to the ex vivo-epithelium (Figure 6A/B). Likewise, beta catenin could be visualized near the cellular membrane on cell culture and tissue sections (Figure 6C/D). In the tissue samples beta catenin expression is detectable in all layers of the porcine cervical epithelium, but the intensity of the immunostaining is lower in the superficial cell layer. This difference could not be detected in the in vitro-model.

To show cervix specific function mucopolysaccharides were visualised. The multilayered cell cultures showed a blue signal in the apical layers after alcian blue staining comparable to the in vivo situation (Figure 6G/H).

We tested different methods to isolate epithelial cells from porcine cervical tissue. Standard growth media and a range of additive-compositions were investigated regarding their influence on proliferation and morphology. The cultured cells were characterized by histology, immunocyto-and-histochemistry. Studying differentiation, we cultured tissue explants and cells from passage one on hanging membranes either with access to medium from both basolateral and apical side or in an air-liquid interface with access to medium only from the basolateral side.

Media, antibiotics and serum for cell culture were supplied by Biochrom AG, Berlin, Germany. All other chemicals were obtained from Sigma-Aldrich, St. Louis, USA, unless otherwise indicated.

Porcine uterine cervices were collected in a local slaughterhouse from healthy, 6-month-old, pre-pubertal animals, approximately 15 min after death. Ovaries of all animals only showed small follicles of the same size as a sign for cyclic inactivity. The organs were washed immediately with Dulbecco's PBS.

Conditions of transport were investigated concerning the influence on cell viability after isolation. Cervical tissues were transported to the laboratory either dry or in growth medium containing 10% FCS, fetal calf serum (on ice or at room temperature).

After optimization of transport conditions the cervical tissues were transported in basic growth medium (Medium 199, or after intial media testing: Ham's F-12) containing 10% FCS at room temperature within 2 hours.

For histology, tissues of four animals were transversally cut in 5 mm thick pieces and fixed in chilled Bouin's solution on-site.

To isolate pure epithelial cells from porcine cervical tissue, we applied different previously published enzymatic digestion methods 71324 as well as a modified protocol for cellular outgrowth from primary tissue explants 25. Briefly, cervices were opened longitudinally and rinsed with Dulbecco's PBS. Tissue pieces of the ectocervical mucosa (approximate length 1 cm, width 2 mm and height 1 mm) were cut with scissors. 10 pieces were placed (mucosal side up) in a 25 cm² culture flask with 2 mL of growth medium. After initial media testing, the growth medium consisted of conditioned Ham's F-12, 10% FCS, 10 ng/mL murine EGF (epidermal growth factor), 1 μg/mL porcine insulin, 100 U/mL penicillin, 100 μg/mL streptomycin, 50 μg/mL gentamycin, 1 μg/mL amphotericin B, 10 μg/mL reduced glutathione and 10 μg/mL ascorbic acid as antioxidants. After 48 and 96 h, culture medium was changed. After five days the tissue explants were removed from the culture vessel.

If necessary for further experiments, the cells were passaged (3 mL of Accutase, 20 min incubation at 37°C) after five days of culture.

Cells were cryopreserved using a standard protocol. 5 × 10⁴ cells were resuspended in 500 μL CryoMaxx (PAA Laboratories GmbH, Pasching, Austria) and cooled down to -80°C (1°C/min) before storage in liquid nitrogen.

To proof cell type and purity of the collected cells we applied immunocytochemistry against pan-keratins. Differentiation was shown using an antibody against beta catenin, as part of adhearens junctions. For primary antibodies and protocol details see Table 2.

Either tissue explants or 1 × 10⁵ cervical cells in passage one were grown on Superfrost Plus^® slides (Gerhard Menzel, Glasbearbeitungswerk GmbH & Co. KG, Braunschweig, Germany) until confluence. They were rinsed with Dulbecco's PBS, fixed in ice-cold acetone for 10 min and washed three times in Dulbecco's PBS. Blocking of unspecific binding sites was not necessary. The primary antibody was incubated for 1 h at room temperature (pan-keratins) or at 4°C overnight (beta catenin). Mouse immunoglobulin G fraction or normal rabbit serum (DakoCytomation, Glostrup, Denmark) served as negative control. After three washing steps in PBS-T 0.01%, the secondary antibody (Goat Anti-Rabbit or Anti-Mouse Poly-HRP, Thermo Scientific, Rockford, USA) was incubated for 30 min at room temperature. After three more washing steps, diaminobenzidin (DAB-/Metal Concentrate 10×, Thermo Scientific) was used as chromogen.

To evaluate different media for further use as basic growth medium, we applied a morphology score.

Cervical epithelial cells of 30 animals were cultured (in conventional cell culture plates) as described above in five different growth media containing 10% FCS as well as antibiotic, antimycotic and antioxidant reagents (Table 3). Cell morphology of primary (five days after explantation) and cells in passage one (two days after passaging) was assessed by light microscopy. The morphology was scored concerning proliferation, presence of vacuoles and epitheloid phenotype (Table 4). The highest achievable score of each medium was 240 points (8 points in three categories, 30 animals in total; primary cells after five days of culture: n = 15 animals; cells in passage one after two days of culture: n = 15 animals).

The growth medium with the best results in the initial media testing was further modified. Conditioned growth medium (mimicking stromal-epithelial interaction) was prepared referring to standard protocols 19. In short, 5 × 10⁵ 3T3 Swiss albino embryo fibroblast cells (ATCC, Manassas, USA) were grown in a 75 cm² cell culture flask in 20 mL Ham's F-12, 10% FCS. After 48 h (subconfluence), fresh growth medium was incubated for 24 h on the fibroblasts. The sterile filtered supernatant was stored at -20°C. To prepare conditioned growth medium, one part of the supernatant was diluted in two parts of unconditioned medium.

Influence on cell viability and proliferation of different growth media (for media composition details see Table 3) was monitored using WST-1 Proliferation Assay (Roche Deutschland Holding GmbH, Grenzach-Wyhlen, Germany) according to the manufacturer's guide. In short, 1 × 10⁴ cervical cells of passage one were seeded in 96-well culture dishes. 24 h, 48 h and 72 h after seeding, cells were incubated for 30 min with the reagent and WST-1 turnover was measured photometrically at wavelength 450 nm and a reference wavelength of 630 nm. All measurements were conducted for three animals as triplicates.

Cells were grown on permeable membranes (6-well and 24-well Millicell hanging inserts, PET membrane, Millipore, Temocura, Canada) hanging in tissue culture plates and creating two compartments, a basolateral and an apical one. We applied different cultivation setups: access to complete growth medium from either sides or air-liquid interface for nutrition exclusively from the basolateral side.

1 × 10⁵ cervical epithelial cells (passage one) were seeded in 24-well inserts containing 200 μL growth medium with 1 mL medium in the well. To culture non-passaged cells on membranes, four pieces of cervical mucosa were placed in a 6-well insert containing 500 μL growth medium with 2 mL medium in the well. The culture plates were incubated at 37°C, 5% CO2 in a humid chamber. After 48 and 96 h growth medium was changed. After five days of culture, the mucosal pieces were carefully stripped off the membrane. To achieve an air-liquid interface, growth medium was completely removed from the insert. After three weeks of culture, the membranes were carefully drawn off the plastic inserts and further processed for histology.

For immunohistochemistry, the cervical tissue or cells on membranes (embedded in 1.2% agarose to ensure a cutting ancle of 90°) were fixed in Bouin's solution at 4°C for at least 24 h. After washing in 4% buffered formalin, the samples were dehydrated and embedded in paraffin. Sections of approximately 5 μm were cut, mounted on Superfrost Plus^® slides and dried at 37°C overnight.

Hemalum/eosin staining to identify cellular morphology and alcian blue staining at pH 2.5 to visualize mucus production was performed for both tissue and membrane sections.

To proof the differentiation state of the epithelial cells cultured on membranes, we applied immunohistochemistry using antibodies against keratins and beta catenin (for primary antibodies and protocol details see Table 2).

After deparaffination and rehydratation, sections were washed in Dulbecco's PBS. Antigens were demasked by boiling slides for 2 min and cooling down for 20 min in citrate buffer, pH6. Endogenous peroxidase was blocked in 3% H₂O₂ in methanol (twice, 15 min). Sections were incubated with blocking solution at 37°C in a humid chamber for 1 h. Sections were incubated with primary antibodies at 4°C in humid chamber over night. Mouse immunoglobulin G fraction or normal rabbit serum respectively (DakoCytomation) served as negative control. After washing in PBS-T 0.01% (3 times, 15 min), slides were incubated with goat anti-mouse or anti-rabbit secondary antibody (Thermo scientific) conjugated with horseradish-peroxidase for 30 min. Peroxidase was visualized using diaminobenzidin.