Decellularization and recellularization of the ovary for bioengineering applications; studies in the mouse

Oophorectomy was conducted on 108 female C57BL/6 N mice (Charles River, Sulzfeld, Germany) aged 10- to 20-weeks and resulted in 216 ovaries that were used for the analysis. The animal work was approved by the local animal ethics committee at the University of Gothenburg, Sweden (#114–2014). The isolated ovaries were dissected free from the surrounding tissue and immediately placed in Perfadex (Ex-vivo, Gothenburg, Sweden) and stored at − 20 °C before further processing.

For the decellularization experiments, the ovaries were thawed and randomly divided into three groups that were exposed to one of the three decellularization protocols previously reported [30]. In brief, the ovaries were submerged and agitated in 0.5% SDS for 10 h (protocol 1, P1; n = 63), or 2% SDC for 16 h (protocol 2, P2; n = 63), or a combination of the two detergents (0.5% SDS for 5 h followed by 2% SDC for 8 h; protocol 3, P3; n = 63). All the ovaries were then washed thoroughly in water, then treated for 30 min with a DNase I solution at 37 °C (40 units/ml; Sigma-Aldrich, Stockholm, Sweden). Each ovary was then washed with water for 24 h, then sterilized with 0.1% peracetic acid for 30 min, and washed for 24 h with sterile PBS supplemented with Gibco’s 1% antibiotic-antimycotic (Anti-Anti; penicillin 10,000 U/mL, streptomycin 10,000 μg/mL and fungizone 25 μg/mL; Thermo Fischer Scientific, Gothenburg, Sweden). All the ovaries were then frozen until analysis. A fourth group of normal unprocessed mouse ovaries that had been frozen once in Perfadex (Ex-vivo) was included as a normal control group for the analysis (n = 27).

All tissues that were processed for histology and immunohistochemistry were fixed in formaldehyde for 1 h, then dehydrated in alcohol and embedded in paraffin. Each ovarian tissue was then cut in a microtome into 5 μm sections. Selected midsections of each specimen were fixed on glass slides. Sections were then dewaxed and rehydrated prior to any staining procedure. Hematoxylin and eosin (H&E) staining was performed according to standard procedures. Some sections were processed with the nuclear dye 4′,6-diamidino-2-phenylindole (DAPI) for 1 min for fluorescence labelling of DNA. Excess DAPI was removed by two PBS washes, then slides were coverslipped with mounting media (F6182, Sigma Aldrich, Stockholm, Sweden). Stained sections were visualized and photographed with a Leica microscope (Micromedic AB, Stockholm, Sweden). To provide contrast for visualization of the scaffold in the images with DAPI staining, the auto fluorescence in the green channel was also photographed and included in the pictures.

Quantitative assessments of collagen, glycosaminoglycans (GAGs), and elastin were conducted using the kits from BioColor according to the manufacturer’s instructions (Fastin™ Elastin Assay; Blyscan™ Sulfated Glycosaminglycan Assay; Sircol™ Soluble Collagen Assay; Sircol™ Insoluble Collagen; Carrickfergus, Northern Ireland, UK). Due to the detection limitations of the kits (5 μg for elastin; 0.25 μg for GAGs, 1.0 μg for soluble collagen, and 10.0 μg for insoluble collagen), two ovary samples were pooled for each of the analysis. Therefore, the results are presented as “μg ECM component/two ovaries”. A total of six normal ovaries were used for the analysis of each component, and a total of 12 decellularized ovaries were used per group and per analysis.

To evaluate whether the produced ovarian scaffolds were sterile, non-cytotoxic, and were capable of harboring cells that can be of therapeutic benefit for follicular support, red fluorescent protein (RFP)-labelled mouse bone marrow MSCs (RFP-MSCs; MUBMX-01201 C57BL/6, Cyagen, Santa Clara, CA, USA) were cultured up to the seventh passage under standard culture conditions (DMEM with Glutamax, 10% fetal bovine serum, 1% Antibiotic-Antimycotic, Thermo Fisher Scientific; humidified chamber with 5% CO₂ enriched air). For ovarian scaffold recellularization, one million cells were used to recellularize each ovarian scaffold by five repeated injections of 200,000 cells per injection. Injections were placed to cover the entire structure of the ovary using a 30G needle connected to a 1 ml syringe (VWR International, Stockholm, Sweden). Each recellularized ovarian construct (n = 12 per scaffold group) was then left for 30 min in the incubator to allow for cell attachment. Following this, additional cell culture medium was added and the constructs were incubated individually in a 6-well plate for 14 days. The culture medium was replaced every second or third day depending on the pH-sensitive color change in the culture medium. Nine constructs from each group were then washed in PBS, fixed with paraformaldehyde and embedded in paraffin blocks and histologically processed as described above (for H&E staining). Immunohistochemistry for RFP was also conducted on sections of recellularized ovarian scaffolds. This was necessary since the paraffin embedment negated the biologically active RFP that was expressed by the genetically engineered cells. For antibody staining, antigen retrieval was performed with citrate buffer (pH = 6.0) in a pressure cooker and 10% normal goat serum diluted in PBS containing 0.2% triton-X100 was used for blocking (1 h). The sections were then incubated with the following antibodies for 1 h at room temperature; RFP (1:100; #ab62341, Abcam, Cambridge, UK), cleaved caspase-3 (1:100; asp175, Cell signaling, Leiden, The Netherlands) and Ki67 (1:100; #16667, Abcam). After several washes in PBS, the fluorescent-labelled secondary antibodies Alexa Fluor 488 (1:300; #150077, Abcam) and/or Alexa Fluor 594 (1:300; #150116, Abcam) were added to the sections for 1 h at room temperature. Sections from a processed mouse spleen was used for a positive control for the staining of Ki67 and cleaved caspase-3 (data not shown).

The recellularization efficiency was manually assessed by observation of the scanned H&E stained 5 μm sections. Cells that were within, and on the surface of the ovarian scaffolding structure were counted. Cells located outside the actual scaffolding structure were not counted. The total counting area was measured and the total number of cells per square millimeter was calculated.

Normal, decellularized and recellularized mouse ovaries (n = 3 per group) were processed for scanning electron microscopy (SEM). Due to the absence of cellular lipid membranes in the decellularized tissues, better contrast was obtained in the specimen for SEM after they were treated with osmium tetroxide and thiocarbohydrazide [37]. The treated samples were dehydrated, mounted, sputter coated with 4 nm gold particles and imaged with a Zeiss Gemini-2 SEM 500 (Carl Zeiss AB, Stockholm, Sweden).

All three decellularization methods were found to be effective based on the successful removal of intracellular components and nuclear material. The ovary-specific extracellular scaffolding structure was better preserved with P1 and P2, than with P3 (Fig. 1).

The mean elastin content in the decellularized ovarian tissue was about 60% compared with the original ovarian elastin content prior to decellularization. This reduction was not significant compared with normal control tissue due to large variations between samples in each group. Protocol 1-treatment lead to a significant decrease of sGAGs, and the levels of both insoluble and soluble collagens were significantly reduced by P1 based on multiple group comparison statistics (Kruskal-Wallis test and Dunn’s multiple group comparison) (Fig. 2a-e). When each of the test groups was compared with the normal group by the Mann-Whitney U-test, P2 and P3 were found to significantly reduce the levels of sGAGs and collagen (p < 0.05 vs. the normal group). However, the levels of sGAGs, and to some degree also the soluble collagen levels, were higher after P2 than after P1 when these groups were compared with each other using the Mann-Whitney U-test (sGAGs: p = 0.0022, soluble collagen: p = 0.065).

Finally, the MTT-assay indicated that no cytotoxic remnants were present in any of the three scaffold types (Fig. 2f).

The RFP-labelled MSCs that were used for the recellularization were able to recolonize large areas of the ovarian scaffolds, including the deeper layers (Fig. 3a-f). All the detected cells were double-labeled with RFP and DAPI, demonstrating that they were all recellularized cells, and did not originate from the scaffold donor as a result of an incomplete decellularization. In general, the ovarian cortex exhibited better recellularization than the medulla-region in all scaffold types. Sections stained with Ki67 and cleaved caspase-3 showed that cells within the scaffolds remained proliferative and non-apoptotic in all constructs 14 days after cell seeding (Fig. 3g-i).

Highly detailed topography assessments of all three types of recellularized ovarian scaffolds with SEM showed that the remaining ovarian ECM structure was significantly repopulated on the surface (Fig. 3j, arrows). It was difficult to visualize the cells in cross sectioned recellularized scaffolds with SEM. However, cells were visualized in the hollow and porous structures, including in the deeper scaffold compartments (Fig. 3k, arrows).

The cell density data indicated that there was no significant difference in cell density between the three recellularized ovarian scaffold types (Fig. 4). However, there was a large intra-group variance (evident by the large error bars). For example, one ovarian scaffold from the P2-treated group was extensively recellularized and had an average of 1277 cells/mm², which is a considerable difference compared with the average number for the most successful construct in the P1-derived scaffolds (916 cells/mm²) and the P3-derived scaffolds (541 cells/mm²; Fig. 4). However, when comparing the cell density of the least efficiently recellularized scaffold in each group, no major difference was seen (236–253 cells/mm² for all the three groups). Hence, even if P2 had several more constructs with greater recellularization efficiency than scaffolds produced by P1 or P3, we did not observe a significant difference between the construct groups due to the inconsistency of recellularization efficiency within each of the groups.