Endometriotic tissue fragments are viable after cryopreservation in an ex vivo tissue model recapitulating the fibrotic microenvironment

Endometriosis is the most prevalent benign gynecologic disease with an enormous societal impact. Approximately 190 million women worldwide are suffering from severe (cyclic) pelvic or abdominal pain or subfertility due to endometriosis. The quality of life of most of these patients is seriously impaired ( van Aken et al. , 2017 ). Thereby, endometriosis-related health care costs are similar to other chronic diseases including diabetes or Crohn’s disease ( Simoens et al. , 2012 ; Zondervan et al. , 2020 ). Clinically, endometriosis is subdivided into three separate disease manifestations: peritoneal endometriosis (PER), ovarian endometrioma or endometriotic cysts (OMA), and deep endometriosis (DE). Current therapies mainly focus on symptom management by analgesics or hormonal therapy based on suppression of estrogen activity. One critical problem in the approach of using hormonal therapy is the incompatibility with a wish to conceive. Additionally, surgical resection can be considered. In recent years, there has been an increasing interest in the fibrotic component of endometriotic lesions ( Li et al. , 2016 ; Zhang et al. , 2016 ; Vigano et al. , 2020 ). Fibrosis and myofibroblast contraction may be one of the causes of endometriosis-related pain ( Odagiri et al. , 2009 ; Liu et al. , 2019 ; Yan et al. , 2019 ; Garcia Garcia et al. , 2022 ). In addition, on the cellular level, endometriotic lesions consist largely of fibrosis and myofibroblasts, formed after differentiation of endometrial-like cells ( Zhang et al. , 2016 ; Yan et al. , 2020 ). The myofibroblasts even seem to have a pivotal role in the process leading to the establishment of endometriotic lesions ( Muraoka et al. , 2023 ). Recently, Vigano et al. (2018) proposed to incorporate lesional fibrosis in the histopathologic definition of endometriosis, which was endorsed by Guo (2018) . Myofibroblasts are activated stromal cells pivotal in wound healing in healthy tissues based on their ability to contract and produce extracellular matrix. Normally, myofibroblasts go into apoptosis after the process of wound healing is completed. However, in fibrotic diseases, there is a persistent survival of myofibroblasts partially caused by the stiff microenvironment, which activates pro-survival signaling in these cells. This leads to a self-stimulating feedback loop resulting in an excessive deposition of extracellular matrix proteins ( Huang et al. , 2012 ; Matsuzaki et al. , 2016 ; D’Urso and Kurniawan, 2020 ; Hinz and Lagares, 2020 ). To study the central role of myofibroblasts as the main cellular effector of fibrosis in endometriosis, it is important to develop ex vivo tissue models that mimic the in vivo situation as closely as possible. While tissue slices have been used to recapitulate the physiological tissue context of several other fibrotic diseases ( Paish et al. , 2019 ; Sun et al. , 2023 ), they have not been used to study fibrosis in endometriosis ex vivo . Cryopreservation of whole tissue fragments is a technique to store tissue fragments for later usage ( Rivas Leonel et al. , 2019 ). By using cryopreservation to store endometriosis tissue fragments collected at different points in time, it may be possible to perform experiments in parallel, thereby increasing efficiency and decreasing inter-experimental variation. The use of whole tissue fragments also incorporates the fibrotic microenvironment in our ex vivo model. Additionally, these tissue fragments can be exchanged between different research institutes. Our laboratory has extensive experience and knowledge in the field of cryopreservation of human ovarian cortex fragments, both for clinical and research purposes ( Peek et al. , 2023 ). In ovarian cortex fragments, viability is best preserved by slow freezing in a cryoprotectant medium based on dimethyl sulfoxide (DMSO), although some laboratories also successfully use ethylene glycol (EG) as cryoprotectant ( Rosendahl et al. , 2011 ; Bastings et al. , 2016 ). Different cell and tissue types might require different freezing strategies, as demonstrated by the long-lasting debate about the most optimal freezing protocol for embryo cryopreservation and various medium compositions for non-fertility-related cryopreservation ( Whaley et al. , 2021 ; Casciani et al. , 2023 ). In this study, we assessed the possibility and the best technique of cryopreserving endometriotic tissue. To this end, we analyzed the viability of the tissue and the preservation of the fibrotic phenotype in fragments from different endometriosis subtypes after cryopreservation and thawing using various cryopreservation techniques. Next, we assessed the possibility to use these tissue fragments for experimental ex vivo studies. Our results indicate that the culture of cryopreserved and thawed endometriotic tissue fragments constitutes an appropriate ex vivo model that provides opportunities for effective exploration of novel therapies for endometriosis.

In this study, we used tissue samples from different types of endometriosis from 28 patients. The mean age of the included patients is 35.7 ± 7.8 (years ± SD), 35.8 ± 8.0 for PER, 35.8 ± 8.1 for OMA, and 35.4 ± 7.2 for DE. The BMI of included patients is 25.9 ± 4.6 (kg/m 2 ±SD). Patients were individually labeled and patient-specific details are presented in Supplementary Table S1 . We collected endometriotic tissue from all endometriosis subtypes and cryopreserved and thawed the fragments. Then, we performed a resazurin assay to compare metabolic activity of the tissue fragments between differently cryopreserved and fresh control tissue. We made comparisons between fresh tissue and tissue cryopreserved according to the three cryopreservation methods. We did not detect significant differences in metabolic activity between fresh tissue and cryopreserved tissue ( Fig. 1 ). Next, we compared DMSO-based cryopreservation with EG- and PROH-based cryopreservation. There were no significant differences in metabolic activity between DMSO-based cryopreservation, EG-based cryopreservation, and PROH-based cryopreservation. Results are shown in Fig. 1 and Table 2 . Resazurin assay . Samples were cryopreserved according to the different cryopreservation protocols. After thawing, samples were cultured for 24 h and incubated with resazurin. Fluorescence of the medium was measured, corrected for weight, and normalized to fresh samples of the same subtype. Metabolic activity is presented as relative fluorescence compared to the mean fluorescence of fresh samples. For all three endometriosis subtypes, no differences in metabolic activity between cryopreservation protocols were observed. Fresh (non-cryopreserved) tissue and tissue frozen and thawed five times without cryoprotectant were used as viable or non-viable controls. Error bars represent SD and squares, triangles and circles represent individual values. DMSO, dimethyl sulfoxide; EG, ethylene glycol; PROH, propane-1,2-diol; PER, peritoneal endometriosis; OMA, ovarian endometriotic cyst; DE, deep endometriosis. Results of the resazurin assay. Metabolic activity is presented as fluorescence of supernatant medium corrected for the weight of the sample and normalized to the mean of fresh samples. DMSO, dimethyl sulfoxide; EG, ethylene glycol; PROH, propane-1,2-diol; PER, peritoneal endometriosis; OMA, ovarian endometriotic cyst; DE, deep endometriosis. Because the tissue fragments cryopreserved according to the various protocols did not show differences in metabolic activity and DMSO-based cryopreservation is the most time-efficient and commonly used cryopreservation technique, we decided to use DMSO-based cryopreserved samples for the subsequent experiments in this study. To assess morphological changes and to confirm the presence of endometrial-like epithelium and stromal cells in all samples used for the viability staining experiments, hematoxylin–eosin staining was performed. Endometriosis could be confirmed in all samples used for the viability staining experiments. The epithelial component of the endometriotic lesions was variable in tissue fragments before cryopreservation. The epithelial layer generally consisted of one layer of large, columnar cells in most cases, as seen in the insets in Fig. 2A (PER and DE). In some cases, the epithelial cells were flattened or atrophic directly after surgical removal ( Fig. 2A , OMA). However, after cryopreservation and culture, samples generally showed enhanced flattening of the epithelium ( Fig. 2A , DE). This was seen in the deep lesion presented in Fig. 2A . The stromal cells, including the myofibroblasts, showed no morphological changes after cryopreservation and subsequent culture. Pyknotic cells were barely seen. In the tissue fragments cultured for 72 h, we also did not observe any morphological changes suggestive for loss of cellular viability ( Fig. 2B ). In the culture for 7 days in some areas, the cellular density appeared to be decreased, which may be a consequence of karyolysis, as shown in the PER and OMA samples ( Fig. 2B , PER and OMA). In the DE sample, some pyknotic cells were observed ( Fig. 2B ). Overall, the tissue did not show signs of cell death-associated features after 72 h culture, but it does to a certain extent after 7 days ( Fig. 2B ). Morphology after cryopreservation and culture . ( A ) Hematoxylin–eosin staining of fragments used for viability staining after cryopreservation and controls. ( B ) Hematoxylin–eosin staining of prolonged cultured tissue fragments. Cryopreserved and thawed tissue was cryopreserved and thawed according to the DMSO-based protocol and cultured for 24 h, 72 h/3 days, or 7 days to reveal any tissue damage resulting from the cryopreservation/thawing protocol. Control tissue was fixed in paraformaldehyde directly after surgical removal. After 24 h in peritoneal and ovarian endometriosis, no morphological changes did occur after cryopreservation and in deep endometriosis epithelial cells are flattened, however, there were no obvious signs of apoptosis or necrosis ( A ). In the prolonged cultured fragments, no morphological changes occurred after 3 days in culture. After 7 days, cellular density seems decreased in the PER and OMA samples and some pyknotic cells are present in the DE sample ( B ). PER, peritoneal endometriosis; OMA, ovarian endometriotic cyst; DE, deep endometriosis. Scale bar =100 μm. To assess the cell-type specific viability, we subsequently performed a viability staining assay. We compared cell viability percentages between DMSO-based cryopreserved fragments and fresh tissue fragments ( Fig. 3 ). The viability staining assay showed an overall viability of 91% (CI: 87–95%) viable cells after cryopreservation with DMSO. In fresh tissue, 89% (CI: 83–95%) of cells were viable (MWU-test P  = 0.66). In PER, 92% (CI: 87–96%) of cells were viable after DMSO-based cryopreservation and culture versus 97% (CI: 95–99%) in fresh peritoneal samples (MWU-test P  = 0.31). In OMA, 95% (CI: 91–98%) of cells were viable, compared to 91% (CI: 89–93%) in fresh tissue (MWU-test P  = 0.40). In tissue samples from DE lesions, 88% (CI: 78–98%) of cells were viable after cryopreservation and culture compared to 79% (CI: 69–90%) in fresh tissue (MWU-test P  = 0.29) ( Fig. 3A ). In our non-viable controls, ViVaFix creates a hazy stain rather than individual dead labeled cells due to massive necrosis. Thereby, quantification of dead cells is not possible in our non-viable controls. None of the tissues derived from the three different endometriosis subtypes contained clusters of dead α-SMA positive cells, indicating high-level survival of the myofibroblasts after freezing and thawing ( Fig. 3A ), which is in line with overall cell viability. Viability staining assay . ( A ) Viability staining—representative images. Samples were cryopreserved according to the DMSO protocol or used fresh, cultured for 24 h, and stained with Hoechst 33342, ViVaFix, and α-SMA antibody. Imaging was performed using confocal microscopy. In cryopreserved and fresh samples, α-SMA is detected in several areas (yellow) and a few cells are labeled as dead by ViVaFix (in red). In the non-viable controls, necrosis creates a hazy stain rather than individually stained cells. For each individual panel: Top left—Cyan—Hoechst; Bottom left—Red—ViVaFix; Top right—Yellow—α-Smooth Muscle Actin; Bottom right—Overlay. Scale bar = 200 μm. ( B ) Quantification of viability staining. Hoechst 33342 and ViVaFix-stained cells were counted and percentages of viable cells were calculated. No significant differences between cryopreserved and fresh samples were found for each endometriosis subtype. Error bars represent SD and squares and circles represent individual values. PER, peritoneal endometriosis; OMA, ovarian endometriotic cyst; DE, deep endometriosis; ns, not significant. To assess the possibility if the cryopreserved and thawed tissue fragments could be used as an ex vivo endometriosis model, we performed intervention experiments. First, we assessed the expression of COL1A1 and CEMIP genes after treatment with pirfenidone or metformin in two whole tissue samples. We found that both fibrotic markers were downregulated after treatment with pirfenidone, but not after metformin treatment. After treatment with pirfenidone, the RT-PCR revealed downregulation of both CEMIP and, to a lesser extent, COL1A1 compared to the GAPDH control ( Fig. 4A ). This downregulation was also detected with RT-qPCR. The fold change for COL1A1 was 0.45 and for CEMIP 0.16 compared to untreated tissue ( Fig. 4B ). Surprisingly, in myofibroblasts isolated by LCM, inhibition of fibrotic gene expression was observed as illustrated by downregulation of CEMIP most evident after metformin treatment and only in a lesser extent after treatment with pirfenidone ( Fig. 5 ). RT-PCR analysis of ex vivo treated tissue fragments . Cryopreserved and thawed tissue fragments from two patients were cultured for 72 h: without treatment (condition 1); with 5 mM metformin (condition 2); with 250 μM pirfenidone (condition 3). ( A ) RT-PCR signal of Cell Migration Inducing Protein (CEMIP) is less intense compared to reference gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH) in both pirfenidone treated (condition 3) patient samples compared to untreated tissue (condition 1). ( B ) RT-qPCR analysis of the same samples. Fold changes of RT-qPCR are calculated using the ΔΔCt-method. The fold change for COL1A1 was 0.45 and for CEMIP 0.16 compared to untreated tissue. Lines represent means and circles and squares represent individual values. RT-PCR analysis of myofibroblasts isolated from cryopreserved and thawed, ex vivo -treated tissue fragments . Sections were cut from one tissue fragments after 72-h culture: without treatment (condition 1); with 5 mM metformin (condition 2); with 250 μM pirfenidone (condition 3). ( A ) Areas rich in myofibroblasts were selected (left panel, yellow dotted line) and isolated by laser capture microdissection (right panel, after laser cutting). Arrowheads point out endometriotic glands. Note that individual cells are difficult to distinguish due to the thickness of the sections. Scale bar = 100 μm. ( B ) RT-PCR signal of Cell Migration Inducing Protein (CEMIP) is less intense compared to reference gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH) in metformin-treated (condition 2) tissue compared to untreated tissue (condition 1).

Between July 2021 and April 2023, endometriotic tissue samples were collected from patients who underwent surgery for endometriosis in our academic medical center. Informed consent was obtained prior to surgery. Because all tissues are left-over material, oral informed consent was sufficient to use the material for research purposes. The local medical ethics committee was consulted about these regulations and confirmed the exemption for further ethical assessment. Tissue fragments from patients aged 18 years and older and with different subtypes of endometriosis were included. In total, we used 65 tissue samples from 28 patients. For the metabolic assay, 30 cryopreserved samples were used, four per subtype for DMSO-based cryopreservation, four (OMA and DE) or three (PER) for EG, and three (OMA) or two (peritoneal and DE) for propane-1,2-diol (PROH)-based cryopreservation. Additionally, three non-cryopreserved samples per subtype were used as controls. For the viability staining and morphology, 12 cryopreserved samples were used, four per endometriosis subtype and additionally three fresh samples per subtype were used as controls. For the prolonged culture, three samples were used, one for each subtype. Two samples were used for the methodological proof-of-concept experiment. Tissue suspicious for endometriosis was removed during surgery with minimal cauterization damage by gynecologists experienced in endometriosis surgery (AN, others are acknowledged). After removal, fragments were divided into two parts. One part was used for the regular diagnostic process, wherein a pathologist confirmed the presence of endometriosis in all cases. The other part was directly placed in ice-cold Leibovitz’s L-15 medium (L-15, Capricorn Scientific, Germany) and transported to the laboratory for direct further processing. Visible cauterization effects and non-endometriotic tissue were removed and the tissue was cut into thin fragments of approximately 5 mm  × 5 mm × 1 mm. Cryopreservation was performed according to three different protocols, all previously used successfully for cryopreservation of ovarian cortex fragments in a clinical setting ( Gook et al. , 1993 ; Rosendahl et al. , 2011 ; Bastings et al. , 2016 ). For all protocols, the fragments were placed in 1 ml of ice-cold medium in a cryotube and incubated on a roller bank. The cryoprotective solutions were prepared according to the compositions shown in Table 1 and filter sterilized. After incubation in cryoprotectant, the cryotubes were placed in a Freezecontrol ® (Freezecontrol ® CL-3300, TL-13 and CryoGenesis™, CryoLogic, Blackburn, Australia) installation for computer-controlled freezing. Temperature was lowered according to the protocols as shown in Table 1 . After slow freezing, cryotubes were stored in liquid nitrogen. Cryopreservation procedures. 2°C/min to −9°C, seeded, checked after 10 min 2°C/min to −9°C, seeded, checked after 10 min 2°C/min to −8°C, seeded, direct continuation of freezing L-15 2% FBS 0.25 mol/l sucrose PBS 0.5 mol/l EG 0.25 mol/l sucrose PBS 1.0 mol/l PROH 0.2 mol/l sucrose L-15 2% FBS 0.1 mol/l sucrose PBS 0.25 mol/l sucrose PBS 0.5 mol/l PROH 0.1 mol/l sucrose L-15 2% FBS PBS PBS Hams F10 2.5 g/l serum albumin Slow freezing and thawing protocols. Percentages in volume/volume. PBS, phosphate-buffered saline; FBS, fetal bovine serum; L-15, Leibovitz-15. Cryopreserved endometriotic tissue fragments were thawed prior to performing the viability assessment assays. Details of thawing protocols are shown in Table 1 . After thawing, fragments were cultured for 24 h on a shaking table at 37°C in humidified air with 5% CO 2 in Dulbecco’s Modified Eagle Medium (DMEM-HA, Capricorn Scientific, Germany), 10% v/v fetal bovine serum (FBS, Heat inactivated fetal bovine serum, 10500-064, Gibco, USA), and 40 µg/ml gentamicin (Centrafarm, Etten-Leur, The Netherlands). During this 24 h culture period, the eventual cell dead induced by the cryopreservation and thawing procedure will become detectable. To assess cell survival based on metabolic activity, a resazurin assay was performed. The non-fluorescent resazurin is reduced into the fluorescent resorufin by mitochondrial activity, whereby an increased fluorescence acts as a proxy for cell viability. After thawing, tissue fragments were cut into smaller fragments of ∼1 mm 3 . As negative controls, fragments were snap-frozen and thawed five times in medium without cryoprotectant to induce cell death. The fragments were cultured 3–5 h in DMEM, 10% FBS, and 40 µg/ml gentamicin. After 3–5 h, medium was changed for medium with 0.1 mg/ml resazurin (Resazurin sodium salt, R7017, Sigma-Aldrich, Germany). After 24-h culture in resazurin, tissue fragments were pat dried and weighed. The fluorescence of the medium was measured with 525–540 nm excitation and 620–640 nm emission light on a Victor™ Platereader (Victor™ X3, PerkinElmer, Waltham, USA). Fluorescence levels were corrected for the weight of the tissue samples and were normalized to the mean fluorescence of the fresh controls per series of experiments and per subtype. To assess individual cell viability in the thawed endometriotic tissue fragments, a staining assay based on cell membrane integrity was performed using VivaFix™ (ViVaFix™ 583/603, No. 1351117, Bio-Rad, Hercules, USA) and Hoechst 33342™ (Hoechst 33342, H3570, Invitrogen, Life Technologies, USA). To this end, thawed tissue fragments were cut into smaller fragments of ∼1 mm 3 . As positive controls, cell death was induced by snap freezing and thawing thrice in medium without cryoprotectant. The fragments were cultured for 24 h in DMEM, 10% FBS, 40 µg/ml gentamicin, Hoechst (1:10 000), and ViVaFix (1:200). After 24 h, fragments were washed thrice in phosphate-buffered saline (PBS, P38135 , Sigma Aldrich, Germany) and fixed for 30 min in paraformaldehyde (PFA, buffered formaldehyde 4%, 4078-9001, Klinipath, The Netherlands). Fragments were washed again thrice for 5 min in PBS, then stained for 24 h with α-smooth muscle actin monoclonal rabbit antibody (α-SMA, 1:200, D4K9N, Cell Signaling Technologies, USA). Secondary antibody staining was done with Alexa 488 monoclonal goat antibody (1:200, Alexa Fluor 488, goat-anti-rabbit, A32731, Thermo Fisher Scientific, USA) for 24 h. For negative controls, staining was done with secondary antibodies only. After washing, fragments were optically cleared in increasing concentrations of fructose solutions, 2 h in 29%, 5 h in 57.5%, and overnight in 115%, all with 0.5% α-thioglycerol to prevent the Maillard reaction ( Palacio-Castaneda et al. , 2021 ). For every fragment, at least five fields were imaged using a Leica SP8 confocal microscope (Leica SP8 SMD) and image analysis was done using Fiji (Fiji is just ImageJ for Mac OS X, version 2.9.0) ( Schindelin et al. , 2012 ). Image processing parameters were adjusted individually per sample because of variation in optical properties between samples. For every image, the total number of cells was automatically counted based on Hoechst staining. The number of dead cells was automatically counted based on colocalized staining of Hoechst and ViVaFix. In this way, the percentage of viable cells was calculated. After imaging, the fragments were embedded in paraffin and 4 μm sections were cut with a microtome. Sections were mounted on glass slides using Stick-On (VWR Chemicals, France) and standard hematoxylin–eosin staining was performed. Additionally, immunohistochemical staining was done to confirm the presence and assess the morphology of endometrial-like glands, stroma, and myofibroblasts. For immunohistochemistry staining, 4 μm-thick slices were mounted on FLEX IHC mounting slides (FLEX K8020, DAKO, Agilent, USA). Sections were deparaffinized in xylene and rehydrated in ethanol before incubation with peroxide (hydrogen peroxide solution 30%, Boomlab, The Netherlands) to block endogenous peroxidase activity for 10 min. Antigen retrieval was done for 10 min in boiling citrate (Citrate Buffer, ScyTek, USA) with 0.5% (v/v) Tween-20 (Tween ® -20, 655204, Merck, Germany). Afterward, sections were cooled down on ice and incubated first with avidin block (Avidin/Biotin Blocking Kit, Life Technologies, USA), washed in PBS, and secondly with biotin block, both for 15 min before washing again. Thereafter, sections were pre-incubated for 10 min with normal horse or goat serum (normal horse serum, S-2000, Vector Laboratories, USA; normal goat serum, S-1000, Vector Laboratories, USA) diluted in PBS containing 0.22% of bovine serum albumin (BSA, 22% in saline, A7034, Sigma Aldrich, Germany). Sections were incubated for 1 h with primary antibodies for cytokeratin-7 (CK7, monoclonal mouse anti-human antibody, diluted 1:200, OV-TL 12/30, Dako) or α-SMA (monoclonal rabbit anti-human antibody, diluted 1:200) at room temperature. After washing in PBS, sections were incubated with biotinylated secondary antibodies (biotinylated goat anti-rabbit IgG, Vector Laboratories, USA; biotinylated horse anti-mouse IgG, Vector Laboratories, USA). After washing in PBS and incubation with ABC complex (ABC Peroxidase staining Kit, Thermo Fisher Scientific, USA) for 40 min at room temperature, sections were stained with 3,3′-diaminobenzidine (DAB, EnVision Substrate Buffer and DAB+, Dako, USA) for 10 min and counterstained with Mayer’s hematoxylin for 90 s before washing in tap water. After the staining protocol, sections were dehydrated in alcohol and xylene baths and covered with Quick-D mounting medium (7281, Klinipath, VWR, The Netherlands) and a coverslip. To test the possibility if our ex vivo tissue fragments could be used for in vitro intervention studies, a proof-of-concept experiment was performed. After thawing, tissue fragments were cut into small fragments of ∼1 mm 3 . Next, the fragments were cultured for 72 h in DMEM, 10% FBS, and 40 µg/ml gentamicin, with or without the addition of 250 μM pirfenidone (Pirfenidone CRS PHRS, Merck, Germany) or 5 mM metformin (Metformin 317240-5GM, Merck, Germany). After 72 h of treatment, tissue samples were divided in half for further analysis. Treated samples and controls were assessed for gene expression analysis using RT-qPCR, as described below. Additionally, untreated samples of all subtypes were maintained in culture for 7 days. These 7 days cultured untreated tissue fragments and 3 days cultured untreated tissue samples were fixed in PFA, embedded in paraffin, and cut into 4 μm-thick slices. After mounting on a glass slide, they were stained with a standard hematoxylin–eosin stain to assess morphologic changes after prolonged culture. To isolate a specific cell population from the heterogeneous endometriotic lesions before subsequent analysis, a laser capture microdissection (LMD) microscope was used. One-half of the treated tissue samples was snap-frozen in liquid nitrogen. Cryosections of 20 µm thickness were cut on a cryostat (Slee MEV+, Adamas, The Netherlands) at −20°C and placed on PEN-membrane slides (PEN-membrane slides 2.0 μm, Leica, Germany). The slides were dried in a desiccator and hematoxylin–eosin stained. Next, areas rich in myofibroblasts were visually selected and cut out using an LMD (LMD6, Leica, Germany). A total of 20 fields of each ∼25 000 µm 2 were collected per sample to reach a sufficient amount of tissue for subsequent RNA analysis. The second half of the ex vivo tissue culture samples as described above was minced and subsequently dissolved in Trizol (Tri-RNA Reagent, Farvorgen, Taiwan). LMD-isolated myofibroblasts were collected and directly dissolved in Trizol. RNA was isolated according to the manufacturer’s instructions and resuspended in nuclease-free water. RNA was quantified using the Qubit 4 Fluorometer and the Qubit RNA BR Assay kit (Invitrogen, USA). cDNA was synthesized by Superscript II (Invitrogen, USA) and random hexamer primers (Promega, USA) using 1 μg of total RNA. Both PCR and RT-qPCR were performed. The PCR reactions were performed in a T100 thermocycler (Biorad, The Netherlands) in 25 μl containing 2 μl of cDNA, and 2 μl of each of the following mixed forward and reverse primers from a 10-μM stock: Glyceraldehyde 3-phosphate dehydrogenase ( GAPDH , FWD 5′-TGGGTGTGAACCATGAGAAG-3′, REV 5′-AGTTGTCATGGATGACCTTGG-3′, Collagen 1A1 ( COL1A1 , FWD 5′-GAGAGCATGACCGATGGATT-3′, REV 5′-CGCTGTTCTTGCAGTGGTAG-3′), and Cell migration-inducing and hyaluronan-binding protein ( CEMIP , FWD 5′-AGCAAACACTTCCTGCACCT-3′, REV 5′-CAGAGCCTCGATGTCCATGAT-3′). The PCR was run for 10 min at 95°C, followed by 35 cycles of 92°C for 1 min (denaturation), 65°C for 1 min for GAPDH and COL1A1 or 51°C for 1 min for CEMIP (annealing) and 2 min at 72°C (extension). The products were run on a 1.5% agarose gel and photographed. For the quantitative PCR (RT-qPCR), a reaction mix was prepared containing 6.25 μl iQ SYBR Green Supermix (BioRad, The Netherlands), 0.5 μl of each forward and reverse primer ( GADPH , hydroxymethylbilane synthase ( HMBS , FWD 5′-TTCTTCTCCAGGGCATGTTC-3′, REV 5′-CTGGTAACGGCAATGCGGCT-3′), COL1A1 and CEMIP ) from a 10-μM stock, 2.75 μl water and 2.5 μl 20-fold diluted cDNA. The qPCR reaction was performed in triplicate in a 96-well plate in a CFX Connect Real-Time PCR detection system (BioRad, The Netherlands) for 7 min at 95°C, followed by 15 s at 95°C for denaturation and 1 min at 60°C for annealing and extension, followed by 39 cycles of 15 s at 95°C and 1 min at 60°C, an additional step of 10 s at 95°C, and 5-s interval increases of 0.5°C from 60°C to 95°C to obtain a melting curve. Fold changes for the genes of interest ( COL1A1 , CEMIP ) were calculated using the ΔΔCt-method using the geometric mean of the reference genes ( GAPDH , HMBS ). Statistical analysis was performed using GraphPad (GraphPad Prism, version 9.5.0). In the metabolic assay, the fluorescence levels were corrected for the background fluorescence and the weight of the samples. The fluorescence was normalized to the mean of the fluorescence of the fresh samples per series of experiments, so metabolic activity is presented as relative activity normalized to fresh tissue samples of its own subtype. A Mann–Whitney U -test (MWU-test) was used to compare relative metabolic activity and cell viability percentages between samples. Values of P  < 0.05 are considered statistically significant.

In this study, we showed that endometriotic tissue fragments are viable after cryopreservation. We performed three independent but complementary assays to assess the viability and metabolic activity of whole endometriotic tissue fragments after cryopreservation and culture. Next, we demonstrated the possibility to use these cryopreserved tissue fragments as an efficient ex vivo endometriosis model. The use of cryopreserved endometriotic tissue fragments mimicking the in vivo situation is likely to benefit research on endometriosis. In contrast to many other experimental in vitro endometriosis models, tissue fragments resemble the in vivo situation closely. The fragments represent the complex and heterogenic cellular and extracellular nature of endometriosis. In addition, cryopreservation enables storage of viable tissue fragments and thereby increases efficiency of research and creates novel research opportunities. By cryopreservation, tissue fragments can be stored and thawed at a self-chosen moment. In this way, experiments can be performed in parallel, which decreases inter-experimental variation. In addition, experiments can be performed during regular working hours and tissue samples could be exchanged between different research institutes. Currently, both 2D and 3D cell cultures and animal models are commonly used to study endometriosis ( Burns et al. , 2021 ; Gołąbek-Grenda and Olejnik, 2022 ). Although co-culturing systems and even more advanced 3D and chip-culture systems are under investigation ( Chen et al. , 2012 ; Boretto et al. , 2019 ), the complexity of the heterogenic endometriotic cellular environment is not fully represented. The various cell–cell interactions between endometriotic cells, healthy surrounding cells, and infiltrated immune cells have not yet been incorporated in one single model. Besides this, the stiffness of the matrix in cell culture models strongly influences the expression of fibrotic markers ( Matsuzaki et al. , 2016 ). Animal models overcome a part of the difficulties that are described above, but they also have strong disadvantages: it is difficult to reliably mimic endometriosis in animal models, they are expensive and may be considered as morally reprehensible. Inducing experimental endometriosis is necessary for these animal models, mostly leading to superficial cystic lesions on the abdominal wall, not recapitulating the human clinical subtypes of peritoneal, OMA and DE ( Greaves et al. , 2014 ; Escudero-Lara et al. , 2020 ). With the use of tissue fragments as an ex vivo model, we approach the in vivo situation more closely than other preclinical models currently used. Cryopreservation is successfully used for storage of other types of tissue. The technique is routinely used for fertility preservation. In this case, ovarian cortex fragments are cryopreserved and stored for periods up to decades. After re-implantation, these fragments are viable and even the ovarian follicles containing the vulnerable oocytes, are able to mature and may lead to normal pregnancies ( Donnez et al. , 2004 ; Dolmans et al. , 2021 ). Recently, it was shown that cryopreservation and transplantation of whole rat organs is another promising step forward in this fast-developing domain of research ( Han et al. , 2023 ). Pirfenidone is a drug known from other fibrotic diseases and is approved by the European Medicine Agency as therapy for idiopathic lung fibrosis patients. It has been shown that pirfenidone downregulates CEMIP expression, a gene involved in epithelial-to-mesenchymal transition and fibroblast-to-myofibroblast transdifferentiation, and other fibrotic markers, both in vivo and in ex vivo fibrosis models ( Kwapiszewska et al. , 2018 ; Lehmann et al. , 2018 ; Liu et al. , 2023 ). In our study, pirfenidone treatment also showed downregulation of markers for fibrosis COL1A1 and CEMIP in the cryopreserved ex vivo tissue fragments. Metformin, classically known for its widespread use as an anti-diabetic drug, is currently also receiving extensive attention due to its anti-inflammatory and potential anti-fibrotic properties. In this light, metformin is studied in various models for fibrotic diseases, endometriosis, and wound healing. Metformin seems to have both anti- and pro-fibrotic effects depending on the tissue type and specific context, which could explain the differences between the effect on whole tissue fragments versus isolated myofibroblasts in our experiments ( Kheirollahi et al. , 2019 ; Huang et al. , 2022 ; Tombulturk et al. , 2024 ). To our knowledge, this is the first study assessing viability and usability of endometriotic tissue fragments after cryopreservation to establish an ex vivo endometriosis model. A strength is our comprehensive combination of experiments to assess tissue viability after cryopreservation and thawing. We have compared three different ways of cryopreservation and so we are confident that viability is preserved well using the DMSO-based cryopreservation protocol. Besides this, we tested all three different endometriosis sub-types and we compared viability between fresh and cryopreserved tissue. Moreover, we did not only investigate the viability based on three different assays, but we also showed the possibility to use the cryopreserved tissue fragments for ex vivo drug intervention experiments. The extensive experience of our laboratory with cryopreservation of ovarian tissue fragments for both clinical and research purposes was valuable in optimizing the cryopreservation and viability assessment of endometriotic tissue samples after thawing. As we found no significant differences in metabolic activity between the different cryopreservation protocols, we concluded that the DMSO-based method is the most suitable technique for future experiments because this technique is the most time-efficient and well-known in other applications. In our proof-of-concept experiment, we showed the successful use of the cryopreserved tissue fragments as ex vivo endometriosis model. We were able to detect gene expression differences both in the heterogenous cell population of the whole tissue samples, as well as in isolated myofibroblasts. Our cryopreserved tissue is usable for various assays, which will be of considerable benefit for future research. Especially the isolation of specific cell types using LMD can be very helpful to tackle the difficulty of the heterogeneity of endometriotic lesions. This study has some limitations as well. Although the viability of the endometriotic tissue after cryopreservation is high in general, there is some variability between tissue samples. However, since this variation was observed both in cryopreserved as well as in fresh tissue, a causal relationship with cryopreservation seems unplausible. A more suitable explanation is a difference in the extent of damage the different tissue samples sustained during the surgical procedure. Although the gynecologic surgeons were instructed to remove the lesion with as little cauterization damage as possible, it is impossible to excise lesions completely without any damage, especially in small (peritoneal) lesions. These variations also complicate any subgroup analyses for example for age or medication use, as these would not reach statistical significance due to the small subgroup numbers. Although the combination of three different experiments assessing various viability-related parameters gives us a representative overview of overall viability, the individual assays have their limitations. Because of the strong staining by resazurin, it is not possible to perform histological analysis on the samples used for the metabolic assay to confirm the presence of endometriosis. However, we did confirm the presence of endometriosis in adjacent tissue fragments that were used for the viability staining assay. In the viability staining assay, a reliable quantification of dead cells in our non-viable control was not possible, because ViVaFix does not stain individual cells in case of massive cell necrosis. The results from the intervention experiments showed decreased gene expression of fibrotic markers after treatment with various agents. The main goal of these experiments was to show the methodological possibilities of our technique and therefore we used a small sample size. The results of these proof-of-concept experiments emphasize the promise of future studies in this direction to robustly evaluate potential anti-fibrotic therapies for endometriosis. Our ex vivo model based on cryopreserved tissue slices facilitates testing potential therapeutics for endometriosis. By using previously stored, cryopreserved tissue, it is more convenient to perform experiments with viable tissue fragments from multiple patients at a time. This increase in numbers of samples will increase the reliability of experiments. With more accurate pre-clinical data, the shift to human in vivo testing will be accelerated. This will hopefully lead to the discovery of new non-hormonal and non-invasive interventions for endometriosis patients in the near future. Recently, several potential therapeutics based on regression of fibrosis, such as relaxin-2 and sodium tanshinone IIA, have been studied in preclinical endometriosis models ( Luo et al. , 2020 ; Yoshino et al. , 2020 ). Because the fibrotic phenotype of endometriosis is preserved in our tissue fragments model, we are able to assess the effect of these and other anti-fibrotic agents ex vivo in human tissue. When treating fibrosis, it is not only essential to stop the formation of fibrosis but also to reduce or regress the fibrosis which is already present in the lesions at the start of therapy. Because a serious delay between onset of the disease and a diagnosis is still very common, fibrosis will have had the opportunity to form over up to several years before therapy could be started ( Staal et al. , 2016 ). Therefore, fibrosis is likely to be extensive at the start of anti-fibrotic therapy in many patients . Immune cells, especially macrophages, are important in both fibrogenesis and fibrolysis ( Duffield et al. , 2013 ; Sindrilaru and Scharffetter-Kochanek, 2013 ; Duan et al. , 2018 ; Hogg et al. , 2020 ). Therefore, assessing the natural presence and phenotype of immune cells in our tissue fragments could be a future research direction. To conclude, our model facilitates the testing of new potential therapeutics in endometriosis. Because our assays demonstrate the possibility to use cryopreserved whole tissue fragment as a novel preclinical endometriosis model, it can provide essential information about the effect of potential therapeutics. In this way, hypotheses about targeting fibrosis in endometriosis can be tested in a reliable and efficient ex vivo system.