Organotypic culture as a research and preclinical model to study uterine leiomyomas

Organotypic cultures of tissue slices have been successfully established in lung, prostate, colon, gastric and breast cancer among other malignancies, but until now an ex vivo model based on tissue slices has not been established for uterine leiomyoma. In the present study, we describe a method for culturing tumour slides onto an alginate scaffold. Morphological integrity of tissue slices was maintained for up to 7 days of culture, with cells expressing desmin, estrogen and progesterone receptors. Driver mutations were present in the ex vivo slices at all-time points analyzed. Cultivated tumour slices responded to ovarian hormones stimulation upregulating the expression of genes involved in leiomyoma pathogenesis. This tissue model preserves extracellular matrix, cellular diversity and genetic background simulating more in-vivo-like situations. As a novelty, this platform allows encapsulation of microspheres containing drugs that can be tested on the ex vivo tumour slices. After optimizing drug release rates, microspheres would then be directly tested in animal models through local injection.

However, this model neither capture three dimensional tumour architecture nor many of the important signalling dynamics of the interactions among all cells present in the tumour and between these and the microenvironment. In addition, rapid disappearance of cells carrying MED12 mutation and decrease of estrogen and progesterone receptors expression as the days of culture progress have been observed, challenging the usefulness of 2D culture model [15][16][17][18] .
Organotypic cultures consist of sectioned tumour tissue into thin slices, mounted onto porous membranes for mechanical support and incubated in a controlled condition 19 . They retain histological and three-dimensional structure (3D), with inter-and extracellular interactions, cell matrix components, and intact metabolic capacity. This approach has been successfully used to gain insights into tumour biology and as a preclinical model for drug discoveries in many different cancers 19,20 .
Our present study aimed to develop a well-characterized 3D organotypic culture system using precision-cut slices from human uterine leiomyoma placed onto an alginate scaffold. In order to determine whether the freshly prepared explants are capable of capturing and then maintaining essential features of the original tumour, tissue slices were harvested at different time points and compared to the original tumour using histology and immunohistochemistry (IHQ). In addition, tumour slices were stimulated with ovarian steroids and selected transcripts and proteins were quantified by real time PCR (qPCR) and western blot, respectively.

Results
Alginate polymer constitutes a suitable scaffold for tissue culture explant. Tissue slices were cultured on 1 mm thickness alginate scaffold which allow preservation of 3D tissue structure in culture and position the tissue slice at the air/liquid interface enabling efficient oxygenation (Supplementary Figure S1). The high porosity of the alginate sponge (average 84%) with interconnected pore network 21 enables the solutes diffusion through the scaffold allowing adequate supply of nutrients and oxygen to the cells. In addition, microsphere-incorporating scaffolds provide appropriated mechanical properties for in vitro culture. In our system, we obtained microspheres diameter mean volume of 59.31 µm (10% < 26.75 µm and 90% < 84.23 µm), with capacity for high drug loading yield and a tunable sustained release 22 , which would be ready for fast testing in animal models after the in vitro assay 23,24 . Morphology is preserved in organotypic tissue cultures. Slices from tumours were maintained in culture for up to 3 weeks and tumour cells morphology was assessed by H&E staining. Morphological integrity of tissues, defined as preservation of general architecture was confirmed in the tumour samples up to 7-10 days of culture. Typical UL histology was observed in the original tumour and derived tissue culture slices, with the presence of areas of smooth muscle fascicles arranged in different orientations surrounded by variable degrees of ECM (Fig. 1). After 1 week of culture, a progressive decrease in cellularity with a concomitant increase in ECM was observed.
Abundant immunoreactivity (-ir) for the muscle marker desmin was found in T0 and the cultured tumour slices (Fig. 1). Similarly, we detected estrogen and progesterone receptors at all-time points analyzed, with a clear signal observed in the nuclei of the smooth muscle cells (Fig. 1). We analyzed the number of immunoreactive cells across different time points (T0, T2, T7, T10) in 3 tumours that preserved tissue morphology for up to 10 days as assessed by H&E staining (Supplementary Figure S2). In one of the tumours DES-ir, ESR1-ir and PGR-ir decreased significantly after 10 days of culture. Significant decreased in ESR1 and PGR signal was also detected after 10 days of culture in tumour L6 and L9, respectively (Supplementary Figure S2). According to these results, ex vivo slices of fibroids could be kept in culture for up to 7 days. Driver mutations are maintained in organotypic tissue cultures. We sequenced the exon 2 hot-spot region of MED12 in the 11 UL analyzed and found mutation in 8 tumours (Supplementary Table S1). DNA sequencing showed that mutations were heterozygous consisting in seven single base substitutions and one in-frame deletion. cDNA sequencing demonstrated that the mutant allele was predominantly expressed in T0 and after 10 days of culture in both point mutation and deletion (Fig. 2a).
Of the tumours employed for long term culture, relative quantification of HMGA2 mRNA in original tumour samples demonstrated upregulation in two tumours (L5, L10) compared to matched myometrium, while tumour L8 showed increased levels in myometrium (Fig. 2b). These 3 tumours increased considerably HMGA2 mRNA levels, 3.5-, 76-and 57-fold after 7-10 days of culture compared to T0 (Fig. 2b). The remaining 3 tumours showed absent or weak expression of HMGA2 mRNA in T0 and during culture. cultivated tumour slices responded to ovarian hormones stimulation across time. In order to determine whether tumour slices responded to ovarian steroids, ex vivo slices from 5 tumours were culture in presence of estradiol, progesterone or a combination of both during 24, 48 and 72 hours and gene-expression profiling of 8 selected hormone responsive genes previously shown to be important for UL pathogenesis were analyzed by qPCR (Fig. 3, Table 1). We observed that tumour slices responded significantly to hormone stimulation increasing expression of selected genes at 24 h, 48 h and 72 h. (Fig. 3). In general, steroids-induced gene transcription was more evident at 24 and 72 hours compared to 48 hours. Interestingly, marked upregulation of HMGA2 was observed with E2, P4 or combinations of both in the tumours analyzed at 24 hours. According to driver alterations, MED12 was mutated in 4 tumours (MED12 mut ) while HMGA2 upregulation was detected in tumour L23 (Supplementary Table S1). When gene expression data from both tumour types were compared, particular expression patterns were observed (Table 1). Thus, E2 stimulated the expression of few genes in MED12 mut , particularly IGF1 and PGR that showed high expression at all-time points analyzed. However, higher number of genes and stronger response to E2 was observed in L23 slices (Table 1). Progesterone upregulated the expression of many genes in both groups at 24 h and 72 h, but the combined effect of both hormones was more evident in L23 slices compared to MED12 mut group.
At protein level, we analyzed PCNA and BCL2 as protein markers for tumour proliferation and anti-apoptosis, respectively. Although at 24 h a slightly inhibitory effect on PCNA expression was observed after hormones addition, particularly with E2 + P4 (p = 0.02), a significant upregulation was observed after 72 hours with either combination ( Fig. 4). At transcription level, we observed that Cyclin D1 (CCND1) whose activity is required for cell cycle G1/S transition, was significantly upregulated by E2 (p = 0.04) and close to significance by P4 (p = 0.06) and E2 + P4 (0.06) at 24 h and its transcript levels seem to remain upregulated at 48 h and 72 h (Fig. 3). E2 (p = 0.03) and P4 (p = 0.02) increased BCL2 protein expression after 24 hours of stimulus, with the same trend observed when both hormones were added (Fig. 4). This expression pattern was similar to that observed for mRNA, with significant upregulation observed after P4 (p = 0.02) and E2 + P4 (p = 0.0039) addition at 24 h.
Epidermal growth factor (EGF) has been demonstrated to play a crucial role as a local growth factor in regulating leiomyoma growth 25 . EGF showed a tendency to increase expression with all the ovarian steroids combination at 24 h, reaching significance at 72 h with E2 (p = 0.046) and E2 + P4 (p = 0.034).
Previous studies have demonstrated that estradiol induces expression of progesterone receptors (PGR) in leiomyoma cells 26 . At mRNA level, we found that E2 or E2 + P4 markedly induced the expression of PGR at all-time points analyzed (Fig. 3). From 48 h to 72 h, PGR mRNA levels were also upregulated after P4 stimulus. At protein level, single or combination of both hormones increased expression of both isoforms after 48 h and 72 h (Fig. 4).
It is worth noting the strong stimulatory effect of sex steroids on cultured slices in genes clearly involved in fibroids pathogenesis, such as VEGFA, COL1A1 and IGF, the latter one being induced by E2 or E2 + P4 from 24 to 72 hours (Fig. 3).

Discussion
We describe for the first time an optimized ex vivo organotypic culture of uterine leiomyoma. This approach has been successfully used in different cancers with tissue morphology and viability preservation for up to 5-7 days 19,27,28 . Here we show that culture of thin tissue explants from UL tumours maintain cell morphology for up to 7 days, and then a progressive decrease in cellularity, hormone receptors and increase in ECM was observed. In addition, slices cultured during 5 days responded to ovarian hormone stimulation regulating the expression of several genes at mRNA and protein level.
Since UL are firm, compact and very fibrotic tumours, it may be difficult to obtain thin sections from them consistently. In our experience, best results were achieved by cutting manually pieces of ~ 0.5-cm-wide × www.nature.com/scientificreports www.nature.com/scientificreports/ 0.5-cm-height before mounting them on the sectioning pedestal. In addition, a carbon steel blade with an angle of 26° was crucial to obtain complete slices and prevent the moving blade to dislodge the entire tissue piece.
Although tissue slices are usually placed on synthetic membrane inserts in 6-well plates commercially available, microspheres containing drugs can be integrated into scaffolds composed of organic materials, such as  www.nature.com/scientificreports www.nature.com/scientificreports/ media and essential for tumour growth, or loss of extracellular matrix and hence all the interactions among ECM and the tumoural cells 18 . However, in the organotypic culture we observed that both, point mutation and in-frame deletion alleles of MED12, preferentially expressed in original tumour, are maintained in derived cultures slices. Since we have used standard media previously used in 2D cultures, our results support the idea that ECM is fundamental for in vitro survival and growth of tumour cells carrying MED12 mutations. In addition, organotypic culture may preserve the rate of fibroblasts to mutant smooth muscle cells, bypassing the weak adherence of SMC to plastic surface and hence avoiding its progressive loss during culture 14,18 .
We have observed that HMGA2 mRNA progressively increased over time in cultured slices suggesting that the organotypic system enhance HMGA2 driver alteration. In addition, we have seen for the first time that ovarian steroids increased HMGA2 mRNA levels in the 5 UL tumours analyzed. High Mobility Group A (HMGA) genes encode DNA-binding non-histone proteins that control cell growth by indirect regulation of the DNA transcription process. The expression of HMGA2 is widely restricted to the embryonic stage and decreases to undetectable levels in adult tissues. HMGA2 expression in normal myometrial cells induces its transformation toward putative tumour-initiating leiomyoma cells 29 . In addition, disturbed expression in adult tissues has an important role in the growth of several mesenchymal tumours 30,31 . Based on our finding, we propose a new mechanism for leiomyogenesis where ovarian steroids induce, directly or indirectly, the expression of HMGA2 promoting LM cell proliferation and growth.
At the transcriptional level, tumours with MED12 mutation barely respond to E2 stimulation (Table 1, Fig. 3) contrary to what was observed in slices from tumour without MED12 mutations. Since MED12 is a component of the mediator complex involved in transcription regulation, mutations in this gene may impair transcription response following ovarian steroids stimulation. In fact, it has been shown that MED12 bind to ESR1 gene promoter and together with cohesin promotes the expression of ERα 32 . MED12 silencing reduced PolII occupancy at the ESR1 promoter and ESR1 mRNA expression. In our dataset, we observed that estrogen increase ESR1 expression at 24 h and 72 h in L23 but not in MED12 mut (Table 1). Interestingly, after 72 hours of E2 + P4 stimulus, a significant decrease of ESR1 mRNA expression was observed in MED12 mut (Table 1). These data strongly support that MED12 plays a major role in ESR1 expression and that mutations in this gene may impair estrogen-regulated transcription. In the hormonal study, we have only one tumour without MED12 mutation, which is in agreement with the high frequency of alterations observed for this gene in UL (50-80%). Therefore, further studies are needed before any conclusion can be drawn.
Integrative genome-scale studies of UL harboring different driver genetic alterations, including those in MED12 and HMGA2 contributing to 80-90% of leiomyomas, have uncovered UL subtypes gene expression signatures 8 . In this study we have observed variable response to hormones stimulation in tumours with two different driver alterations, supporting that UL organotypic cultures maintain molecular mechanisms to respond differentially according to its genetic background. This heterogeneity in tumour response to hormones stimulation may reflect the remarkable variable response observed in patients, with UL showing increase, decrease or not change in size to progestin stimulation 33 . In conclusion, organotypic cultures of UL preserve particular local conditions, such as EMC, cellular diversity, genetic background and autocrine/paracrine signals constituting a more biologically representative system   and 2019 were enrolled in this study. Tumours analyzed included 7 intramurals, 1 submucous and 3 subserous tumours, up to 4-12 cm in size, as well as the matched myometrial tissue. Myometrial samples were taken as far away as possible from leiomyomata. All patients were Caucasian and underwent hysterectomy for menorrhagia without any previous hormonal treatment for at least 3 months (Supplementary Table S1). The histopathological analysis using standard H&E staining and performed by a pathologist indicated tumours with benign histology with no sign of malignancy, nuclear atypia, mitotic figures or necrosis. After surgery, samples were immediately submerged in sterile Hank's balanced salt solution (HBSS) supplemented with 0.25 µg/mL amphotericin B, 100 U/ mL penicillin and 100 µg/mL streptomycin (Sigma-Aldrich Co., St. Louis, MO, USA). Tissue pieces were then transported to the laboratory and processed under sterile conditions. Samples were processed within 4 hours after surgery. organotypic tissue culture. Alginate scaffold disc of approximately 1 mm thickness and 6 mm of diameter were added in groups of 2-4 per well in six-well plates. Tissue slices of approximately 5 ×5 mm were loaded onto the scaffolds and 1 mL of Dulbecco's modified Eagle's medium (DMEM, Biowest, France) supplemented with 10% fetal bovine serum (FBS, Lonza, Spain), 100 U/mL penicillin, 100 µg/mL streptomycin and 2 mM L-Glutamine (Sigma-Aldrich) was added to each well. A droplet of medium was added on top of the tissue slice to create a thin film of liquid that helps to maintain the explant humid 20 . Tissue culture was performed at 37 °C in a 5% CO 2 humidified incubator. Medium was changed every day. Tissue slices from 6 tumours were cultured on scaffolds during 3 weeks to assess for cell morphology, IHC staining and analysis of driver alterations throughout the culture period. For 5 tumours, we obtained 36 tissue slices for sex steroid stimulation at 24 hours (12 slices) 48 hours (12 slices) and 72 hours (12 slices). Three biological replicates were analyzed for each condition: vehicle, estrogen, progesterone and estrogen plus progesterone.
Tissue slices from tumours prior to culture (T0) and from each time point, were submerged in RNAlater and formalin for protein and nucleic acid isolation and for histology and immunohistochemistry (IHC) analysis, respectively. ovarian hormones stimulation. To abrogate the menstrual cycle-dependent influence on the biological characteristics of cells, slices were cultured 2 days in phenol red-free DMEM (Lonza) supplemented with 10% charcoal-treated FBS (Lonza), 100 U/mL penicillin, 100 µg/mL streptomycin and 2 mM L-Glutamine. On the third day, tissue explants were treated with 3.67 × 10 −8 M of 17β-Estradiol (E2, Sigma-Aldrich), 3.18 × 10 −7 M of Progesterone (P4, Sigma Aldrich), or combination of both (E2 + P4) and vehicle (unstimulated control) for 24, 48 and 72 hours. The concentrations of sex steroids used were within the physiological tissue concentrations noted in leiomyoma and myometrium 38 . Steroids were initially prepared as stock solutions at 10 −3 M in absolute ethanol and were stored at −20 °C. Working dilutions were prepared in phenol red-free DMEM medium immediately at the start of the experiment and were stored in ice. All cultures including the controls contained equal amounts of the vehicle, with the final ethanol concentration being 0.002%. nucleic acid and protein isolation. To avoid degradation, tissue sections were immersed in RNAlater and kept at 4 °C until processed. Tissue slice was placed on lysis matrix D (MP Biomedicals, Europe) containing Scientific RepoRtS | (2020) 10:5212 | https://doi.org/10.1038/s41598-020-62158-w www.nature.com/scientificreports www.nature.com/scientificreports/ 500 µL RNAtidy G (PanReac AppliChem ITW companies, Darmstadt) and homogenized using FastPrep-24 TM 5 G instrument (MP Biomedicals) twice for 30 s with a speed of 6 m/s. The sample tubes were placed on ice for 5 min between pulses. Then, chloroform was added to the lysate, centrifuged and the top aqueous phase was precipitated with isopropanol to isolate RNA according to the manufacturer's instructions. To the remaining solution, 100% ethanol (3:10 with RNAtidy G) was added, centrifuged 5 min at 2000 × g and the phenol-ethanol phase was moved into a new tube for protein isolation. The pellet containing DNA was washed 2x with 500 µL of sodium citrate 0.1 M/10% ethanol and resuspended in TE. To the protein extract, 2 volumes of isopropanol were added, washed 2x with ethanol 95%, dried and resuspended in lysis buffer (20 mM EDTA, 140 mM NaCl, 5% SDS, 100 mM Tris pH 8.0 and protease inhibitors cocktail, Roche) at 50 °C for 1-3 hours until their complete solubilization 39 .
Residual genomic DNA was removed by incubating the RNA samples with RNase free DNase I and RNasin (Promega Corp. Madison, WI, USA) according to the manufacturer's instructions. The effectiveness of the DNase treatment was assessed in samples with no reverse transcriptase added (RT-negative). Integrity of RNA was checked by agarose gel electrophoresis. Finally, RNA was quantified by absorbance using a NanoDrop ND-1000 spectrophotometer (ThermoFisher Scientific, Waltham, MA, USA).
MED12 mutation detection. Amplification of MED12 DNA was performed with primers located in intron 1 and intron 2 covering the hot spot region where 99% of mutations have been described (Supplementary  Table S2). To check for the presence of mutations in cDNA, a primer pair located in exon 1 and exon 2 was used 40  The PCR products were separated by agarose gel electrophoresis, and the amplicon sizes were verified by comparison with a 100-bp DNA ladder.
PCR products were cleaned up by ExoSAP treatment using Illustra ™ ExoProStar ™ 1-Step (GE Healthcare Biosciences, Pittsburgh, PA, USA) following the manufacturer's instructions. Sequencing reactions were performed for both strands with the same primers used in the amplification at the Genomic Service of the University of La Laguna (SEGAI).
For each sample forward and reverse electropherograms were checked manually, edited and assembled using MEGA6 41 . Sequence alignment was performed using CLUSTAL W 42 as implemented in MEGA6. Plates data were imported into the GenEx ver 6.1.1.550 data analysis software (MultiD Analyses AB) and interpolate calibration with 3 samples were used to normalize the plates to eliminate run-to-run variation. Data were represented as individual triplicate runs and as averages triplicates (with outliers excluded). Grubbs test was used to remove outliers from each group. The relative expression of each gene was calculated using the 2 −ΔΔCt method normalized to expression levels of two reference genes (GNB2L and PUM1) stably expressed 43  Western blots. Protein concentration was determined using Micro BCA TM protein assay kit (ThermoFisher Scientific). 20 µg of total protein extracts were loaded on 12% sodium dodecyl sulfate-polyacrilamide gel electrophoresis (SDS-PAGE) and then transferred onto polyvinyldifluoride membranes (Bio-Rad, Spain) for 90 min at 350 mA constant. Membranes were blocked with 5% non-fat dry milk in TBS (100 mM Tris, 0,9% NaCl, pH 7,5) containing 0,1% Tween 20 (TBS-T) during 1 h. The membrane was incubated with one of the following mouse monoclonal anti-human antibodies: anti-BCL2 (Santa Cruz Biothecnology, sc-7382; 1:300), anti-PCNA (Santa Cruz Biotechnology, sc-56; 1:100), anti-EGF (Santa Cruz Biotechnology, sc-1667799; 1:200), anti-PGR (Santa Cruz Biotechnology, sc-166169; 1:200), and anti-β-actin (Santa Cruz Biotechnology, sc-47778; 1:1200), overnight in a wet chamber at 4 °C. After three washes in TBS-T, membranes were incubated with mouse-IgGk-BP-HRP secondary antibody diluted 1:2000 (sc-516102, Santa Cruz Biotechnology) for 1 h at 4 °C. After washing, protein bands were developed using an ECL chemiluminescence detection kit (Western bright ECL, Advansta Inc., Menlo Park, CA, USA) and automated image capture of blots was achieved using ChemiDoc XRS + system (Bio-Rad).

Quantitative polymerase chain reaction (qpcR). Retro
Densitometric analysis of protein bands was performed by using Image J program, version 1.52a software (NIH). Integrated optical density for each band was calculated, and then each band was normalized with the integrated optical density of the corresponding β-actin band. Fold difference between relative density units of treated and untreated samples were then calculated. (2020) 10:5212 | https://doi.org/10.1038/s41598-020-62158-w www.nature.com/scientificreports www.nature.com/scientificreports/ Histology and immunohistochemistry evaluation. Leiomyoma and matched myometrium samples were fixed in 10% buffered formalin, then embedded in paraffin and cut in 3-μm-thick sections. The sections were deparaffinated, hydrated and hematoxylin-eosin (H&E) stained to assess the morphological integrity of tissue samples.
Quantitative analysis of the immunoreactive cells. ESR1, PGR and DES immunoreactive cells per microscopic field were counted using computer-based image analysis software (ImageJ, NIH, Bethesda, MD). Five randomly selected fields were digitized on each slide, and then the immunoreactive cells were counted at 400 magnifications. Depending on the immunoreaction, cell profiles with immunostained cytoplasm and unstained nucleus or cell profiles with immunostained nucleus and unstained cytoplasm were counted in each case. Further, stained fragments and artifacts were excluded by the program itself, applying a threshold for minimum object size. Intensity of immunolabeling, evaluated in arbitrary units of gray levels ranging from 1 (black) to 256 (white), was used as criterion for cell immunoreactivity assessment. Cells were considered reactive with values above the ones of their background control sections.
Statistical analysis. GraphPad Prism v. 5.0 (GraphPad Software, La Jolla, CA, USA) was used for all statistical analysis. A one-way repeated measures ANOVA was used for cell counting across time, followed by Dunnett's multiple comparison test to compare means for different times to that of the baseline group (T0). qPCR and western blot data were analysed using paired t-test (one-tailed P-value) after assessing that datasets passed normality tests. A one-tailed Wilcoxon matched-pairs signed-ranks test was used when dataset did not pass normality test. For all analysis, P < 0.05 was considered to be statistically significant.

Data availability
The data that support the findings of this study are available from the corresponding authors on reasonable request.