Acta Biomaterialia 10 (2014) 5034–5042

Contents lists available at ScienceDirect

Acta Biomaterialia journal homepage: www.elsevier.com/locate/actabiomat

Towards the development of a bioengineered uterus: Comparison of different protocols for rat uterus decellularization M. Hellström a,⇑, R.R. El-Akouri a, C. Sihlbom b, B.M. Olsson b, J. Lengqvist b, H. Bäckdahl e, B.R. Johansson c, M. Olausson d, S. Sumitran-Holgersson d, M. Brännström a a Laboratory for Transplantation and Regenerative Medicine, Institute of Clinical Sciences, Department of Obstetrics and Gynecology, Sahlgrenska Academy, University of Gothenburg, Sweden b Proteomics Core Facility, University of Gothenburg, Gothenburg, Sweden c Electron Microscopy Unit, Institute of Biomedicine, University of Gothenburg, Gothenburg, Sweden d Department of Surgery, Institute of Clinical Sciences, Sahlgrenska Academy, University of Gothenburg, Sweden e Chemistry Materials and Surfaces, SP Technical Research Institute of Sweden, Borås, Sweden

a r t i c l e

i n f o

Article history: Received 16 April 2014 Received in revised form 17 July 2014 Accepted 18 August 2014 Available online 25 August 2014 Keywords: ECM (extracellular matrix) Scaffold Transplantation Tissue engineering Infertility

a b s t r a c t Uterus transplantation (UTx) may be the only possible curative treatment for absolute uterine factor infertility, which affects 1 in every 500 females of fertile age. We recently presented the 6-month results from the first clinical UTx trial, describing nine live-donor procedures. This routine involves complicated surgery and requires potentially harmful immune suppression to prevent rejection. However, tissue engineering applications using biomaterials and stem cells may replace the need for a live donor, and could prevent the required immunosuppressive treatment. To investigate the basic aspects of this, we developed a novel whole-uterus scaffold design for uterus tissue engineering experiments in the rat. Decellularization was achieved by perfusion of detergents and ionic solutions. The remaining matrix and its biochemical and mechanical properties were quantitatively compared from using three different protocols. The constructs were further compared with native uterus tissue composition. Perfusion with Triton X-100/dimethyl sulfoxide/H2O led to a compact, weaker scaffold that showed evidence of a compromised matrix organization. Sodium deoxycholate/dH2O perfusion gave rise to a porous scaffold that structurally and mechanically resembled native uterus better. An innovative combination of two proteomic analyses revealed higher fibronectin and versican content in these porous scaffolds, which may explain the improved scaffold organization. Together with other important protocol-dependent differences, our results can contribute to the development of improved decellularization protocols for assorted organs. Furthermore, our study shows the first available data on decellularized whole uterus, and creates new opportunities for numerous in vitro and in vivo whole-uterus tissue engineering applications. Ó 2014 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

1. Introduction Infertility affects 15% of all couples and in most cases this infertility can be circumvented by assisted reproductive techniques (ARTs) such as in vitro fertilization. Despite the great advances in ART, one major subgroup of female infertility remains untreatable. This absolute uterine factor infertility (AUFI) is due to the absence of a functional uterus [1]. The only option for these women to gain genetic motherhood is with the assistance of a gestational surrogate mother. However, this procedure is not approved in most parts of the world for ethical, religious or legal

⇑ Corresponding author. Tel.: +46 721 875141; fax: +46 31 7411701. E-mail address: [email protected] (M. Hellström). http://dx.doi.org/10.1016/j.actbio.2014.08.018 1742-7061/Ó 2014 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

reasons. Uterus transplantation (UTx) has been proposed as a possible curative treatment for AUFI [2], and so far 11 human UTx attempts have been made, with the majority of cases arising from a recent clinical study from our group [3]. However, no viable pregnancy has yet been reported. Although, human allogeneic UTx may prove to be successful in terms of pregnancy, there will still be problems related to the surgical risk to retrieve the uterus from a live donor and the adverse effects that are related to the intake of immunosuppression [4]. Recent progress in stem cell research and biomaterials has given hope to the replacement of a live donor as an organ source, with a customized organ construct for transplantation using autologous cells [5], thus circumventing main obstacles in organ transplantation such as lack of donor organs and long-term immunosuppressive treatments. Successful animal studies have shown that

M. Hellström et al. / Acta Biomaterialia 10 (2014) 5034–5042

complicated organs such as the heart [6], the liver [7], the lung [8,9] and the kidney [10] can be decellularized and recellularized with new cells to construct a tissue engineered organ with at least partial organ specific functionality. The less complex, hollow tissues were recently reconstructed for the clinic; decellularized trachea [11] and blood vessel [12] were recellularized with autologous stem cells and transplanted to patients with success in terms of restored function. Various techniques for decellularization and recellularization have been examined for pre-clinical interventions of miscellaneous tissues, e.g., small intestine [13,14], urinary bladder [15], heart valves [16], larynx [17], dermis [18], musculofascial tissue [19] and nerves [20,21]. Concerning the female internal genital tract, small patches of myometrial tissue were recently decellularized using a modified protocol that had been optimized for decellularization of blood vessels [22]. Other tissue engineering approaches for uterine tissue include the use of scaffolds of collagen/matrigel [23–25] or silk/collagen [26], as well as scaffolds made from blood clots [27]. Promising results using biodegradable polymer scaffolds for ‘‘subtotal uterine tissue replacement’’ in a rabbit model was recently reported in a review [28]. It should be pointed out that the above-mentioned uterus-related tissue engineering studies were not aimed at constructing a complete organ that could replace a uterine graft at UTx. Furthermore, decellularized scaffolds created from whole organ vascular perfusion protocols provide several advantages over decellularized tissue patches or artificial scaffolds: (a) it comprises the correct tissue-specific extracellular matrix (ECM) composition, (b) it has the physical three-dimensional (3-D) appearance of the original organ with the correct conduit architecture for blood vessels and glandular ducts, (c) cells can be reintroduced and sustained in a bioreactor via the vascular conduits for ex vivo organ regeneration, (d) the ECM can be remodeled by repopulated cells and (e) the construct provides guiding cues for tissue-specific cell differentiation and migration [29–31]. The present study aimed to develop a scaffold-design standard for whole uterus which later can be used for recellularization and transplantation studies in rodents. Specifically, we compared three different protocols: two were based on the detergent Triton X-100 and the high ionic solution dimethyl sulfoxide (DMSO) as cell membrane disrupting agents; the third protocol was based on sodium deoxycholate (SDC), which disrupts cells by both detergent and ionic mechanisms. These scaffolds may also become valuable for novel uterine-specific 3-D cell culturing systems to study early embryo implantation and development and endometrium cancer cell behavior or for various in vitro drug screening applications.

2. Materials and methods 2.1. Animal work and uterus isolation In total 35 female Lewis rats (140–180 g; Charles River, Sulzfelt, Germany) were used as whole-uterus donors. Eight of these organs were used as a comparative group for normal tissue, and the remaining 27 uteri were divided into three groups (n = 9 per group) and were exposed to various decellularization protocols. All animal work was approved by the Animal Ethics Committee in Gothenburg, Sweden. The uterus isolation surgery was performed aseptically under isoflurane anesthesia. A laparotomy was performed via a mid-line incision from the symphysis pubis to the xiphoid process. To gain full exposure to the uterus and its vascular tree, the small intestines were initially retracted to the left side of the animal and packed in moistened gauze. The inferior mesenteric artery was cauterized using bipolar diathermy (Coa-Comp Bikoagulator; Instrumenta AB, Billdal, Sweden) at a maximum effect of 2 W.

5035

The colon was divided just above the rectum and retracted to the side to obtain a free operating field. In the subsequent order, the inferior epigastric vessels, the external iliac vessels and the external pudendal vessels were double ligated and cut. All vessels supplying the lower portion of the cervix and the bladder (superior vesical, inferior vesical, cervical and vaginal vessels) were then closed to prevent bleeding when the cervix was transected. Special caution was taken to avoid the vascular branches from the hypogastric trunk on both sides. The cervix was gently lifted up to expose the intact uterine artery and veins, which also enabled the attachment of a titanium clip (Weck Closure Systems Ltd, Resarch Triangle Park, NC, USA) on the superior gluteal vessels. The common iliac vessels and the lumbar vessels (2–4) were then transected between double ligatures. Both uterine horns were then double ligated and cut at a site between the tip of the uterine horn and the oviduct. The uterine horns were then dissected free from their attachments to the dorsal peritoneum and folded over to expose the hypogastric trunk. The aorta and the vena cava were tied off inferior to the ovarian vessels by 8–0 suture to create a complete segment that included the inferior aorta, the inferior vena cava and the intact vascular connection to both uterine horns. Small incisions were made in the vena cava and in the aorta so that the specimen could be gently flushed with 1–2 ml ice-cold Perfadex (+4 °C) solution (Vitrolife Sweden AB, Kungsbacka, Sweden) supplemented with xylocaine (0.4 mg ml–1; Astra Zeneca, Mölndal, Sweden) and heparin (50 IU ml–1; Leo Pharma AB, Malmö, Sweden). The perfusion of the isolated donor uterus was assessed successful when the uterus blanched and clear fluid excited the incision made in the vena cava. The uterus was then placed in the same solution and frozen gradually by first placing it in 20 °C, and then in 80 °C for long-term storage. 2.2. Decellularization of whole uterus and sterility control test The uterus was thawed and the tip of a catheter prefilled with phosphate buffer saline (PBS; pH 7.4) including 0.05% sodium azide (PBS+A) was inserted into the lumen of the aorta and fixed with a ligature. A custom-made decellularization perfusion set-up was built using sterile tissue culture plastics (Supplementary Fig. S.1A) and was connected to a Masterflex perfusion pump (Cole-Parmer instruments, Chicago, USA). The uterus was then perfused with PBS+A for 30 min to remove blood remnants and cell debris. Perfusion speed was set to 6 ml min–1, and the flow was divided into three 23G needle exits, where only one of them was connected to the uterine tissue (Supplementary Fig. S.1B). The remaining two open exits acted as pressure valves, to reduce the risk of possible hypertension within the uterine vasculature during perfusion. 27 uteri were randomly allocated to either of three different protocols for decellularization (n = 9 per group), and the decellularized constructs were compared to native uterus tissue (n = 8). 2.2.1. Protocol 1 (P1; n = 9): the uterus was perfused for 4 h with DMSO (4% in PBS+A) followed by another 4 h perfusion with Triton X-100 (1%, diluted in PBS+A), followed by 30 min perfusion with PBS+A. The uterus was then left in a PBS+A solution overnight at room temperature. The DMSO, Triton X-100 and washing steps were then repeated another four times (for a total five cycles; 5 days). 2.2.2. Protocol 2 (P2; n = 9): the uterus was perfused with five identical cycles as for protocol 1 but with modifications between cycles; all dilutions (including washes and overnight storage) were in dH2O + sodium azide (0.05%; dH2O+A) and a freeze–thaw procedure (80 °C) was introduced between cycles 2 and 3.

5036

M. Hellström et al. / Acta Biomaterialia 10 (2014) 5034–5042

2.2.3. Protocol 3 (P3; n = 9): the uterus was perfused with SDC (2%, diluted in dH2O+A) for 6 h followed by 2 h with dH2O+A. The uterus was then stored in dH2O+A at room temperature overnight. This cycle was repeated four additional times (for a total five cycles; 5 days). After completion of cycle 5 of all protocols, a sterilization process was carried out by perfusing per-acetic acid (0.1% in PBS) through the decellularized tissue for 30 min, followed by sterile PBS for 1 h. The decellularized tissue was then transferred to a new sterile tube and immersed in PBS containing 100 U ml–1 penicillin and 100 lg ml–1 streptomycin at +4 °C for 5 days for an extended washing period. A sterility test was performed on the solution in which the decellularized tissue was stored using standard culture procedures to check for possible aerobic, anaerobic or yeast contaminants. Decellularized tissue was then frozen at 80 °C for long-term storage. All further analyses were made on thawed decellularized organs. 2.3. Histology, immunohistochemistry and electron microscopy The integrity of decellularized uterine blood vessel conduits was analyzed by perfusing a dye through the vascular system (Batson’s #17, Polysciences, Eppelheim, Germany, mixed in cooking oil) from one organ from each decellularization protocol. These organs were not used further for any other analysis, thus the reaming group sizes were n = 8 for each group. A ring-shaped segment from the uterus horn, and one piece from the lower v-shaped region from all the remaining decellularized uteri (n = 24), were formalin-fixed and dehydrated in ethanol/ xylene baths and paraffin-embedded before being cross-sectioned on a microtome (5 lm; HM355S; Thermo Scientific, Cheshire, UK). All sections were rehydrated before further analysis. The general morphology was investigated by light microscopy after hematoxylin and eosin (H&E) staining. DNA was detected with 40 ,6-diamidino-2-phenylindole (DAPI; Life Technologies, Stockholm, Sweden). For immunohistochemistry, antigen retrieval with pronase (0.1%; for fibronectin staining) or citrate buffer (10 mM; pH 6.0; for all other antibodies) was performed before using a Mach 3 polymer detection kit and a Warp Red Chromogen kit (Biocare Medical, CA, USA). All primary antibodies were used at 1:100 dilutions and were from Abcam (Cambridge, UK) unless stated otherwise. Antibodies against collagen I (#ab292), collagen IV (#ab6586), laminin (#ab11575), elastin (#ab23748) and fibronectin (#ab6328) were used to detect ECM molecules. Antibodies against a-smooth muscle actin (1:500, #ab5694) and desmin (1:200; #ab8592) were used to investigate the presence of intracellular proteins after decellularization. The presence of major histocompatibility complexes (MHCs) class I (RT1A/OX-18; Biolegend; #205201, San Diego, USA) and class II (OX-6; #ab23990) were used to examine tissue immunogenicity after decellularization. Stainings with omitted primary antibody on normal uterus tissue acted as a negative control for all antibodies. One tissue sample from each group was also prepared for transmission and scanning electron microscopy using standard procedures. 2.4. DNA and protein quantification Two ring-shaped segments (one for DNA quantification and one for protein quantification) were taken from the right uterus horn from each decellularized (n = 24) and normal uterus (n = 8) and were placed on dry filter paper to remove excess fluid. Each sample was then weighed and homogenized separately with a steel bead in a Tissuelyser (3  20 s at 301/s; Qiagen, Sollentuna, Sweden). A DNA extraction kit (DNeasy Blood & Tissue, Qiagen) was used following the manufacturer’s protocol and the DNA concentration

was established on a Nanodrop (Thermo Scientific, Wilmington, USA). Two separate samples with the highest DNA concentration from each decellularization protocol were loaded onto a 1% agarose gel including GelRed nucleic acid stain (Biotium #41003, Hayward, USA) and were run for 30 min at 80 V. The DNA fragmentation size was compared to a 1 kb DNA ladder. Due to the variation of DNA concentration between samples and groups, and the limited loading volume, 1000 ng of DNA was loaded from P1, 70 ng from P2 and 500 ng from P3. Total protein extraction was performed following the manufacturer’s instructions (Millipore #2140, Merck, Darmstadt, Germany), and the concentration was established using a Coomassie protein assay kit (Thermo Scientific, #23200) and a plate reader. 2.5. Proteomics Proteins were extracted from scaffold and tissue (n = 3 from each decellularized group) by a lysis buffer (8 M urea, 4% 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (w/v), 0.2% sodium dodecyl sulfate (SDS) (w/v), 5 mM ethylenediaminetetraacetic acid and homogenization was performed in a FastPrep apparatus, followed by the addition of SDS to a final concentration of 4%, as described in Supplementary methods information, protein extraction for proteomic analysis. Absolute quantification was performed according to the iBAQ method [32], using UPS2 protein standard spiked into the protein sample followed by 1D-SDS-PAGE separation and digestion of the whole lane of proteins. For relative quantification, tryptic peptides were labeled with tandem mass tags (TMTs) [33], fractionated by strong cation exchange chromatography (SCX) and then separated with reversed-phase nanoLC (C18). We used an Easy-nanoLC system connected to an LTQ-Orbitrap Velos instrument with CID (iBAQ) or stepped HCD fragmentation (TMT) for the QMS analysis. For further details, see Supplementary information. 2.6. Elastin and sulfated glycosaminoglycan (sGAG) quantification Two pieces (one for each quantification method) of decellularized (n = 24) and normal uterus (n = 8) were placed on a dry filter paper to remove excess fluid, and weighed. The ECM-related compounds elastin and sGAG were quantified using colorimetric based assays developed by Biocolor (Fastin™ Elastin Assay; Blyscan™ Glycosaminoglycan Assay; Carrickfergus, UK) using the manufacturer’s instructions and a plate reader. 2.7. Mechanical tests Two to five ring-shaped segments from each decellularized and normal rat uterus were tested regarding their tensile properties with an Instron 5566 (Instron, Norwood, MA, USA) with a pre-load of 0.1 N with a test speed of 20 mm min–1. The accuracy of the tensile tester was 0.5% in force and 0.5% in elongation, based on calibrations performed regularly according to ISO 7500-1:2004 and ISO 9513:1999. Each piece of tissue was cut into a 3 mm wide ring-shaped sample and measured with a vernier caliper (Mitutoyo, Kawasaki, Japan). The samples were threaded onto two U-shaped cylindrical grip hooks and mounted to test radial deformation. Samples from decellularization protocols P1 (n = 22), P2 (n = 23) and P3 (n = 22) were compared to normal (untreated) samples (n = 25). The uterus samples originated from eight different rats for each decellularization treatment. Maximum force and elongation to maximum force were recorded and were normalized to the measured width of the samples. The work required deforming the samples completely and the elastic moduli were also calculated.

M. Hellström et al. / Acta Biomaterialia 10 (2014) 5034–5042

2.8. Statistical analysis Data is given as %, or as means ± standard error of the mean (SEM). Analysis of variance followed by post hoc and Tukey’s test were used for multiple group analysis. Welch’s t-test was used for comparisons of the relative quantitative (TMT) proteomics data. Mann–Whitney U tests were performed to compare results from mechanical studies. Significant levels for all tests: ⁄p < 0.05, ⁄⁄ p < 0.01, ⁄⁄⁄p < 0.001. 3. Results 3.1. Sterility control, morphology and DNA content after decellularization Sterility control tests showed that all decellularized organs were free from bacterial or yeast contaminants (data not shown). Macroscopically, a decellularized rat uterus appeared white, and resembled native uterus in size and shape (Fig. 1A and B). Scaffolds developed by all protocols had preserved vascular conduits, comparable to the architecture of normal uterine blood vessels (Fig. 1C). H&E showed preservation of tissue structure (Fig. 1D–G) and DAPI staining showed substantial removal of DNA (Supplementary Fig. S.2A–D). DNA quantification showed that protocols P2 and P3 removed 99% and 94% of the total DNA respectively (Fig. 2A; on average 33 ± 4.4 ng DNA per mg wet tissue remained after P2, and 161 ± 18.3 ng DNA per mg wet tissue after P3). P1 was significantly less effective in removing DNA with 32% of the DNA still remaining (819 ± 131.9 ng DNA per mg wet tissue). The normal DNA content in an unprocessed rat uterus was measured to 2574 ± 110.5 ng DNA per mg wet tissue. DNA fragmentation after decellularization was most efficient with P2 ( 0.05; ⁄⁄p > 0.01; ⁄⁄⁄p > 0.001.

Table 1 The following proteins were selected from the identified proteins that could only be detected in the P1/P2 (64 proteins in total) and P3/P2 (43 proteins in total) TMT proteomic analyses. Acc #

Protein

Description Available from URL: http://www.uniprot.org 2013 Dec.

P1/P2 unique proteins

P55062 P29457 Q6MGB2 Q6P7Q6 P70482 Q4G024

Bax inhibitor 1 Serpin H1 Procollagen XIa2 Galectin Lamin C2 Collagen 15a1

Suppressor of apoptosis Collagen binding, regulation of proteolysis Extracellular matrix, structural constituent Carbohydrate binding Intermediate filament, structural molecule Cell adhesion, extracellular matrix, structural constituent

P3/P2 unique proteins

Q6IFV3 Q6P6Q2 Q6P725

Keratin type I Keratin type II Desmin

Intermediate filament, structural constituent of cytoskeleton Intermediate filament, structural constituent of cytoskeleton Intermediate filament, structural molecule

For an extended list, see Supplementary Table S.1.

Table 2. Using the relative proportion data from our TMT quantification, and combining it with the iBAQ analysis, we obtained an estimate absolute value for the same proteins located in P1- and P3-generated scaffolds, (Table 2, Supplementary Tables S.1 and S.2). 3.3. Immunohistochemistry and colorimetric based quantification of ECM components Immunohistochemistry staining confirmed the presence of key ECM components arranged in a physiological normal distribution (Supplementary Fig. S.3). Collagen I was evenly dispersed throughout all layers of the uterus, while collagen IV and laminin were typically seen in the myometrium and in endothelial layers around blood vessels and glandular structures. Elastin was predominantly

preserved in the myometrium (Supplementary Fig. S.3I–L), and when quantified, 46% of the total elastin remained in P3-produced scaffolds (8.4 ± 0.6 lg per mg wet tissue), which was significantly more compared to tissue treated with P2 (22%; 3.9 ± 0.5 lg per mg wet tissue). P1 treated tissue had 33% of the total elastin content left (6.0 ± 0.7 lg per mg wet tissue; normal uterine tissue = 18.1 ± 0.9 lg per mg wet tissue; Fig. 3). Furthermore, a large variation between the total sGAG content was seen between the protocols; quantification showed that 52% of sGAGs remained in the tissue after P1 (68.2 ± 14.9 ng per mg wet tissue), which was significantly more than after P2 or P3 (14% and 19%, respectively; P2 = 18.4 ± 2.1 ng per mg wet tissue; P3 = 24.7 ± 5.8 ng per mg wet tissue; normal uterine tissue = 132.2 ± 10.0 ng per mg wet tissue; Fig. 3).

5039

M. Hellström et al. / Acta Biomaterialia 10 (2014) 5034–5042 Table 2 Absolute protein quantification from iBAQ proteomic analysis (values are AVG from n = 2). Acc #

P02454 F1LS40 P13941 F1MA59 F1M6Q3 G3V763 F1LQ00 D3ZUL3 F1LNH3 F1LQC3 D3ZZT9 F1LPD0 F1LR02 P47853 Q01129 P04937 P51886 B5DFC9 Q9EQP5

Protein

Collagen 1a1 Collagen 1a2 Collagen 3a1 Collagen 4a1 Collagen 4a2 Collagen 5a1 Collagen 5a2 Collagen 6a1 Collagen 6a2 Collagen 12a1 Collagen 14a1 Collagen 15a1 Collagen 18a1 Biglycan Decorin Fibronectin Lumican Nidogen-2 Prolargin

ng/mg wet tissue Native (sample 1)

Native (sample 2)

Native (AVG; n = 2)

P1*

P2 (sample 1)

P2 (sample 2)

P2 (AVG; n = 2)

P3*

112.3 68.6 0.0 0.0 0.0 0.0 8.3 639.4 210.8 23.1 83.9 8.0 5.2 50.1 178.2 82.1 156.6 43.9 9.3

58.3 36.9 0.0 0.0 1.6 0.0 2.1 301.5 121.4 11.5 50.8 1.8 14.2 16.4 66.1 57.4 147.6 1.3 6.2

85.3 52.8 0.0 0.0 0.8 0.0 5.2 470.5 166.1 17.3 67.3 4.9 9.7 33.3 122.2 69.7 152.1 22.6 7.7

343.1 225.1 0.4 16.3 13.4 10.1 4.4 5764.7 3868.0 24.2 – 10.9 11.2 111.0 107.8 7.8 50.9 13.7 26.9

144.5 73.4 0.0 0.0 0.0 0.0 8.7 2662.6 1846.1 6.5 0.0 6.4 3.8 50.1 78.4 6.9 53.7 0.6 12.8

607.9 420.0 1.0 23.7 27.5 17.9 0.0 5191.1 3561.4 42.8 7.1 5.9 9.6 109.6 99.2 6.3 4.0 14.1 4.2

376.2 246.7 0.5 11.9 13.8 8.9 4.3 3926.9 2703.8 24.6 3.5 6.1 6.7 79.9 88.8 6.6 28.9 7.3 8.5

529.7 343.3 0.3 11.3 8.7 13.3 5.7 3157.6 2030.3 21.8 – – 6.7 15.2 23.1 9.9 6.8 4.8 4.0

The complete list of quantified proteins is available in Supplementary Table S.2. * Estimated values based on combined TMT/iBAQ analysis.

Fig. 3. The SDC-based decellularization protocol (P3) generated scaffolds with significantly more elastin compared to the Triton X-100/DMSO/H2O based protocol (P2). The PBS-buffered decellularization reagents used in P1 significantly enhanced the remaining levels of sGAGs.

Immunohistochemistry on the intracellular protein smooth muscle cell b-actin (SMA) confirmed its removal. However, desmin staining showed that there were some remnants left in scaffolds produced by P1 and P3, but not after P2 (Supplementary Fig. S.4A–H). The immunogenic proteins MHC class I and II were non-detectable after decellularization using all protocols (Supplementary Fig. S.4I–P). 3.4. Mechanical properties All decellularization protocols affected the extensibility of the samples, which extended further before breaking compared to normal tissue samples (data not shown). Maximum load recordings showed that P2-produced scaffolds needed significantly lower forces to break compared to normal uterus tissue, or compared to P3-produced scaffolds (Fig. 4), and tissue treated with P3 was not different from normal tissue but needed significantly more force to deform compared to P1-produced scaffolds (Fig. 4). Consistently, the total work required to completely deform the tissue showed that the SDC-decellularization protocol (P3) yielded a significantly stronger scaffold compared to the Triton

Fig. 4. Hollow cylindrical samples were stretched radially with a tensile tester and the maximum load, and the work required to completely deform the samples was recorded and normalized to the measured width of the samples. Tissue treated with P2 needed significantly lower forces to break compared to normal uterus tissue and to P3-generated scaffolds. Tissue treated with P3 needed significantly greater force to break than tissue treated with P1. The total work required to completely deform the tissue showed that the SDC-decellularization protocol (P3) yielded a significantly stronger scaffold compared to native tissue and compared to the Triton X-100/DMSO-decellularization protocols (P1 and P2).

X100/DMSO-decellularization protocols (P1 and P2). Albeit a high intra-group discrepancy, the scaffold stiffness (modulus) was unaffected after decellularization (data not shown).

3.5. Electron microscopy analysis Transmission electron microscopy (TEM) revealed fiber structures resembling collagen and elastin in all groups together with abundant gray areas which in normal tissues are GAG-rich (Supplementary Fig. S.5A). Some nuclear cell remains were visualized in tissue treated with P1, with occasional eosinophil granulocytes present (Supplementary Fig. S.5B). P3-generated scaffolds contained dark organized structures around collagen-like bundles, which were more prominent than in other decellularization protocols (Supplementary Fig. S.5C). Cross-sections at low magnification using scanning electron microscopy (SEM) showed abundant fiber structures and few intracellular remnants in any of the scaffolds. However, scaffolds generated from P1 and P2 appeared denser than P3-generated scaffolds (Fig. 5A–D). P3 resulted in a highly organized and porous scaffold that at higher magnification showed

5040

M. Hellström et al. / Acta Biomaterialia 10 (2014) 5034–5042

Fig. 5. (A–D) Cross-sections of normal uterus and of scaffolds from each protocol viewed at low SEM magnification, showing all uterus layers (m, myometrium; e, endometrium; lu, lumen). P3 appeared more porous than P1 and P2 after decellularization. Although all tissues looked rich in fibrous content, the structural difference in fiber organization was obvious at high magnification (E–G). At higher magnification, P3 seemed to have retained its collagen-interlinked connections better (H), which may give rise to its cleaner and more uniform fibrous arrangement.

a cleaner and more uniform fibrous composition (Fig. 5E–G). P3 scaffolds also displayed an arrangement with better preserved inter-linked fiber connections (Fig. 5H) compared to the other scaffolds. Tissues from all protocols showed an acellular perimetrium, and had a luminal surface free from epithelium cells, leaving a basal membrane-like structure.

4. Discussion To our knowledge, the present study is the first to describe the initial steps in the development of a bioengineered whole uterus, which can become beneficial for numerous in vitro studies (e.g. endometrial cancer cell studies and drug screening applications) and for future in vivo transplantation studies with the aim to cure AUFI. We compare three different decellularization protocols and provide a detailed quantitative description of the integrity of the remaining ECM and its biochemical composition following vascular decellularization. Based on earlier successful results from our team, where Triton X-100 and SDC were successfully used for vein, trachea and small intestinal decellularization [12,14,34], we decided to base two of our protocols on repeated cycles with Triton X-100 (P1 and P2) and a third protocol on SDC (P3). Triton X-100 is a detergent that disrupts the cell membrane, but apart from SDC, it is not hypertonic. Therefore, we decided to also add DMSO, which is a high ionic solution and causes cell lysis by osmotic gradients. The only differences between P1 and P2 were that dilutions and washing steps were buffered (PBS) in P1, and non-buffered in P2 (dH2O). One extra freeze–thaw treatment was also introduced in P2. Our third protocol (P3) was based on repeated cycles of SDC and deionized water. SDC is a hypertonic detergent and may therefore be a suitable decellularization reagent. Furthermore, SDC has successfully been used for trachea, larynx and small intestine decellularization and has shown to retain some angiogenic factors in the remaining matrix [14,35–37]. SDC is generally considered to be a less aggressive detergent than the commonly used SDS [29,30,38]. The decellularization perfusion set-up used in our study was custom-built using standard sterile-packed tissue culture plastics. The set-up was low cost, highly reproducible and could maintain organ sterility throughout the decellularization process without using laminar air flow facilities. Based on gross morphological examination, no apparent differences were noted between tissues after the various decellularization protocols. All scaffolds had conduits with similar structure and organization to normal

uterine blood vessels. Vascular leakage was not noted during blood-vessel-fill experiments, which indicated an appropriate vascular pressure during decellularization. Histological analysis showed preserved tissue organization with little cellular material left. However, DAPI staining revealed protocol-dependent differences, and DNA quantification showed that adjustments from using PBS-buffered reagents (P1) to non-buffered reagents (dH2O; P2) during decellularization had a significant impact on the remaining DNA content, which suggests important decellularization properties for dH2O. Based only on the DNA content and considering some general decellularization guidelines [29], P2 in this study gave the best results. It is important to consider the immunogenicity of lingering DNA fragments [29,39–41], but the significance of remaining DNA in scaffolds for tissue engineering is a matter of debate. For example, a recent publication reports no tissue rejection after a heterotopic transplantation of rat aortic root scaffolds which contained 30% of its original DNA content [42], a similar amount to what was left in scaffolds produced by our least effective protocol (P1) in this study. Future in vivo scaffold rejection studies and recellularization experiments will provide more detailed information on the immunological and cell supporting properties of the uterus scaffolds produced in this study. Immunohistochemistry showed that all decellularization protocols used in this study successfully removed the immune reactive elements MHC class I and class II, suggesting that the constructs are suitable for future in vivo applications. Further scaffold composition analysis using TMT proteomic analysis indicated that there were higher amounts of intracellular proteins left in P1-generated scaffolds compared to the other constructs, in particular nuclearrelated proteins. In contrast, the SDC-based protocol (P3) seemed to generate scaffolds with a higher amount of cytosolic or intracellular structural proteins (such as keratins and desmins). We have occasionally observed lasting remnants of intracellular structural proteins in other in-house whole-organ decellularization experiments (unpublished observations), and the impact of these proteins on future recellularization/transplantation attempts remains obscure. However, desmins have been linked to facilitate tissue repair and regeneration and seem to play an important role in cell–matrix interactions [43], whereas keratin was successfully used as a biomaterial component that assisted peripheral nerve regeneration in a rodent model [44]. Many factors that facilitate stem/progenitor cell recruitment have been identified, but numerous mechanisms remain obscure [45]. Several groups report that exposed intracellular components and activated immune cells to some degree play a part in stem cell homing during tissue

M. Hellström et al. / Acta Biomaterialia 10 (2014) 5034–5042

regeneration and wound healing [46]. Thus, some remnants of intracellular proteins or DNA left in the scaffolds may therefore give important cues for autologous stem cell recruitment during scaffold repopulation after transplantation. Maintaining an appropriate level on the immune system may be necessary [47], and even if one of the main objectives with tissue engineered organs is to overcome the need for immunosuppressive treatment after transplantation, clinicians may consider administering a shortterm immunosuppressive therapy while the last remaining immunogens from the scaffold are cleared and replaced with autologous cell/protein components. In order to establish important additional functional information, experiments to evaluate growth factor bioactivity and/or angiogenic properties of the scaffolds will be considered in the future [35–37]. Our proteomics analysis identified additional protocol-dependent differences; for example, P3-generated scaffolds contained higher ratios for the extracellular structural proteins fibronectin (50%) and laminin b1 (Lamb1; 200%) compared to scaffolds produced by P2. Thus, a better preservation of these molecules was achieved with the SDC-based protocol. On the other hand, fibromodulin, which may play a role in the assembly of the ECM, was 240% higher after P1-generated scaffolds compared to P2-generated scaffolds. This indicates that fibromodulin was sensitive to ionic fluctuations and was better preserved with buffered decellularization reagents. Absolute protein quantification with iBAQ proteomic analysis on native uterus tissue and on P2-generated scaffolds showed an enrichment of ECM-related proteins (such as the collagens) and a major reduction in cellular proteins after decellularization. To our knowledge, iBAQ has not previously been used to determine protocol differences and to analyze scaffolds for tissue engineering applications. We found this method particularly good for quantifying the collagens and their different isoforms. Furthermore, by combining iBAQ with conventional TMT quantification ratio data, we were able to estimate absolute values on the collagens and other proteins from the P1- and P3-generated scaffolds. We detected no major difference in collagen content between the scaffolds, and immunohistochemistochemical identification of collagens I and IV confirmed an arrangement comparable to a physiological normal distribution. Collagen is known to be sensitive to decellularization protocols, and original levels usually decrease during decellularization and storage times [48,49] which was the reason we choose to freeze down the finished decellularized uterus scaffolds in 80 °C for long term storage. On the other hand, collagen levels in uterine tissue naturally fluctuate during decidualization, which are the progesterone-dependent changes of the endometrium during the luteal phase. Thus, collagen IV regenerates from only being present around blood vessels and glandular structures during the follicular phase, to be generously found throughout the uterine tissue during the luteal phase when stromal cells rebuild the uterine ECM [50]. It is therefore plausible that assorted ECM molecules lost in the presented decellularization protocols could be restored spontaneously by recellularized cells under the influence of sex hormones [24,50]. The SDC-based protocol (P3) was more efficient in preserving elastin compared to the Triton X-100/DMSO/H2O treatment in P2, suggesting that SDC is an appropriate decellularization reagent for elastin-rich tissues. In contrast, P1 preserved half the original amount of the sGAGs, which was better than other protocols tested, suggesting that sGAGs are sensitive to ionic fluctuations. SEM analysis of the decellularized uterus scaffolds also revealed protocol differences; decellularization with Triton X-100/DMSO (P1 and P2) gave rise to a compact ECM structure, whereas the SDC-based protocol (P3) resulted in a porous scaffold. These observations are comparable to what we have seen in other organs after decellularization (ongoing experiments). These differences may not be important for thin hollow structures such as small intestine

5041

or blood vessels [12,14], but may play a crucial role for thicker, solid organs where constructs need to be repopulated deep into the tissue layers and where a permeable ECM is an advantage. However, scaffold blood vessel conduits remained patent after all our protocols, which opens up the possibility of adding cells via vascular perfusion [6–8,10] that may surpass potential physical barriers for cell migration through the matrix itself. At high electron microscopy magnification (TEM and SEM), individual collagen fibrils in P3-generated scaffolds were observed with a defined organization pattern consisting of repetitive collagen–collagen cross-linked nanofibers. These observations were not seen in the collapsed/compact scaffolds produced by P1 or P2. A similar organization of collagen fibrils (and what may disrupt them) has been described in detail elsewhere [51] and seem to consist of complex proteoglycan chains with proteins from the decorins (or decorans), versicans, laminins and fibronectins [52]. Our TMT proteomic analysis showed that P3-generated constructs contained 100% more versican, and 50% more fibronectin. Together with the high levels of preserved elastin, and despite the reduced decorin content, we believe that this explains why P3 gave rise to a cleaner and more uniform fibrous scaffold arrangement. The mechanical experiments of our study showed that these P3-generated constructs were stronger (maximum load and total work) than P1- and P2-generated constructs. Based solely on maximum force, P2-produced scaffolds even showed a reduced maximum load compared to native tissue, which indicated that P2 impaired some of the original strength and was harsher to the ECM compared to the other protocols. P1, and particularly P3, produced scaffolds that better resembled normal uterine tissue characteristics. However, P3-produced scaffolds required greater force to stretch, and the impact of this physical attribute during a possible pregnancy in a bioengineered uterus is unclear.

5. Conclusions In this study, we developed a scaffold-design standard for future whole uterus tissue engineering requirements which should become useful for numerous applications [25,53]. To our knowledge, the present study shows the first available data on vascular perfusion protocols for the rat uterus, and we conclude that in particular two of the three tested protocols (P2 and P3) are promising candidates for future uterus re-engineering applications. These two protocols proved superior in DNA removal and were in general less damaging to the ECM compared to P1. We additionally present important detailed quantitative protocol-dependent structural and biochemical differences after using PBS-buffered or non-buffered Triton X-100/DMSO decellularization protocols, and also compare them with an SDC-based protocol. These novel findings should aid in the development of improved organ-specific decellularization protocols, not only for the uterus but also for other organs with similar tissue architecture.

Disclosure We report no conflict of interest. Acknowledgements We thank Mrs Ann Wallin, Mrs Yvonne Josefsson and Ms Kanita Cukur for technical aid. This work was funded by VR, LUA and Hjalmar Svensson’s foundation. The EM Unit was supported by the Lundberg Research Foundation. Mass spectrometry analysis was provided by the Proteomics Core Facility at Sahlgrenska Academy, University of Gothenburg.

5042

M. Hellström et al. / Acta Biomaterialia 10 (2014) 5034–5042

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.actbio.2014. 08.018. References [1] Kisu I, Banno K, Mihara M, Suganuma N, Aoki D. Current status of uterus transplantation in primates and issues for clinical application. Fertil Steril 2013;100(1):280–94. [2] Diaz-Garcia C, Johannesson L, Enskog A, Tzakis A, Olausson M, Brannstrom M. Uterine transplantation research: laboratory protocols for clinical application. Mol Hum Reprod 2012;18(2):68–78. [3] Brannstrom M, Johannesson L, Dahm-Kahler P, Enskog A, Molne J, Kvarnstrom N, et al. The first clinical uterus transplantation trial: a six-month report. Fertil Steril 2014. [4] Azimzadeh AM, Lees JR, Ding Y, Bromberg JS. Immunobiology of transplantation: impact on targets for large and small molecules. Clin Pharmacol Ther 2011;90(2):229–42. [5] Badylak SF, Taylor D, Uygun K. Whole-organ tissue engineering: decellularization and recellularization of three-dimensional matrix scaffolds. Annu Rev Biomed Eng 2011;13:27–53. [6] Ott HC, Matthiesen TS, Goh SK, Black LD, Kren SM, Netoff TI, et al. Perfusiondecellularized matrix: using nature’s platform to engineer a bioartificial heart. Nat Med 2008;14(2):213–21. [7] Uygun BE, Soto-Gutierrez A, Yagi H, Izamis ML, Guzzardi MA, Shulman C, et al. Organ reengineering through development of a transplantable recellularized liver graft using decellularized liver matrix. Nat Med 2010;16(7):814–20. [8] Ott HC, Clippinger B, Conrad C, Schuetz C, Pomerantseva I, Ikonomou L, et al. Regeneration and orthotopic transplantation of a bioartificial lung. Nat Med 2010;16(8):927–33. [9] Petersen TH, Calle EA, Zhao L, Lee EJ, Gui L, Raredon MB, et al. Tissueengineered lungs for in vivo implantation. Science 2010;329(5991):538–41. [10] Song JJ, Guyette JP, Gilpin SE, Gonzalez G, Vacanti JP, Ott HC. Regeneration and experimental orthotopic transplantation of a bioengineered kidney. Nat Med 2013;19(5):646–51. [11] Macchiarini P, Jungebluth P, Go T, Asnaghi MA, Rees LE, Cogan TA, et al. Clinical transplantation of a tissue-engineered airway. Lancet 2008;372(9655): 2023–30. [12] Olausson M, Patil PB, Kuna VK, Chougule P, Hernandez N, Methe K, et al. Transplantation of an allogeneic vein bioengineered with autologous stem cells: a proof-of-concept study. Lancet 2012;380(9838):230–7. [13] Totonelli G, Maghsoudlou P, Garriboli M, Riegler J, Orlando G, Burns AJ, et al. A rat decellularized small bowel scaffold that preserves villus-crypt architecture for intestinal regeneration. Biomaterials 2012;33(12):3401–10. [14] Patil PB, Chougule PB, Kumar VK, Almstrom S, Backdahl H, Banerjee D, et al. Recellularization of acellular human small intestine using bone marrow stem cells. Stem Cells Translational Med 2013;2(4):307–15. [15] Brown AL, Brook-Allred TT, Waddell JE, White J, Werkmeister JA, Ramshaw JA, et al. Bladder acellular matrix as a substrate for studying in vitro bladder smooth muscle-urothelial cell interactions. Biomaterials 2005;26(5):529–43. [16] Lichtenberg A, Tudorache I, Cebotari S, Ringes-Lichtenberg S, Sturz G, Hoeffler K, et al. In vitro re-endothelialization of detergent decellularized heart valves under simulated physiological dynamic conditions. Biomaterials 2006;27(23):4221–9. [17] Ma R, Li M, Luo J, Yu H, Sun Y, Cheng S, et al. Structural integrity, ECM components and immunogenicity of decellularized laryngeal scaffold with preserved cartilage. Biomaterials 2013;34(7):1790–8. [18] Hoganson DM, O’Doherty EM, Owens GE, Harilal DO, Goldman SM, Bowley CM, et al. The retention of extracellular matrix proteins and angiogenic and mitogenic cytokines in a decellularized porcine dermis. Biomaterials 2010;31(26):6730–7. [19] Wang L, Johnson JA, Chang DW, Zhang Q. Decellularized musculofascial extracellular matrix for tissue engineering. Biomaterials 2013;34(11): 2641–54. [20] Hu Y, Leaver SG, Plant GW, Hendriks WT, Niclou SP, Verhaagen J, et al. Lentiviral-mediated transfer of CNTF to Schwann cells within reconstructed peripheral nerve grafts enhances adult retinal ganglion cell survival and axonal regeneration. Mol Ther: J Am Soc Gene Ther 2005;11(6):906–15. [21] Godinho MJ, Teh L, Pollett MA, Goodman D, Hodgetts SI, Sweetman I, et al. Immunohistochemical, ultrastructural and functional analysis of axonal regeneration through peripheral nerve grafts containing Schwann cells expressing BDNF, CNTF or NT3. PLoS ONE 2013;8(8):e69987. [22] Young RC, Goloman G. Allo- and xeno-reassembly of human and rat myometrium from cells and scaffolds. Tissue Eng Part A 2013. [23] Lu SH, Wang HB, Liu H, Wang HP, Lin QX, Li DX, et al. Reconstruction of engineered uterine tissues containing smooth muscle layer in collagen/ matrigel scaffold in vitro. Tissue Eng Part A 2009;15(7):1611–8. [24] Wang HB, Lu SH, Lin QX, Feng LX, Li DX, Duan CM, et al. Reconstruction of endometrium in vitro via rabbit uterine endometrial cells expanded by sex steroid. Fertil Steril 2010;93(7):2385–95.

[25] Schutte SC, Taylor RN. A tissue-engineered human endometrial stroma that responds to cues for secretory differentiation, decidualization, and menstruation. Fertil Steril 2012;97(4):997–1003. [26] House M, Sanchez CC, Rice WL, Socrate S, Kaplan DL. Cervical tissue engineering using silk scaffolds and human cervical cells. Tissue Eng Part A 2010;16(6):2101–12. [27] Campbell GR, Turnbull G, Xiang L, Haines M, Armstrong S, Rolfe BE, et al. The peritoneal cavity as a bioreactor for tissue engineering visceral organs: bladder, uterus and vas deferens. J Tissue Eng Regen Med 2008;2(1):50–60. [28] Atala A. Tissue engineering of reproductive tissues and organs. Fertil Steril 2012;98(1):21–9. [29] Crapo PM, Gilbert TW, Badylak SF. An overview of tissue and whole organ decellularization processes. Biomaterials 2011;32(12):3233–43. [30] Gilbert TW, Sellaro TL, Badylak SF. Decellularization of tissues and organs. Biomaterials 2006;27(19):3675–83. [31] Keane TJ, Londono R, Turner NJ, Badylak SF. Consequences of ineffective decellularization of biologic scaffolds on the host response. Biomaterials 2012;33(6):1771–81. [32] Schwanhausser B, Busse D, Li N, Dittmar G, Schuchhardt J, Wolf J, et al. Global quantification of mammalian gene expression control. Nature 2011;473(7347):337–42. [33] Thompson A, Schafer J, Kuhn K, Kienle S, Schwarz J, Schmidt G, et al. Tandem mass tags: a novel quantification strategy for comparative analysis of complex protein mixtures by MS/MS. Anal Chem 2003;75(8):1895–904. [34] Berg M, Ejnell H, Kovacs A, Nayakawde N, Patil PB, Joshi M, et al. Replacement of a tracheal stenosis with a tissue-engineered human trachea using autologous stem cells: a case report. Tissue Eng Part A 2014;20(1–2):389–97. [35] Baiguera S, Gonfiotti A, Jaus M, Comin CE, Paglierani M, Del Gaudio C, et al. Development of bioengineered human larynx. Biomaterials 2011;32(19):4433–42. [36] Baiguera S, Jungebluth P, Burns A, Mavilia C, Haag J, De Coppi P, et al. Tissue engineered human tracheas for in vivo implantation. Biomaterials 2010;31(34):8931–8. [37] Haag J, Baiguera S, Jungebluth P, Barale D, Del Gaudio C, Castiglione F, et al. Biomechanical and angiogenic properties of tissue-engineered rat trachea using genipin cross-linked decellularized tissue. Biomaterials 2012;33(3): 780–9. [38] Reing JE, Brown BN, Daly KA, Freund JM, Gilbert TW, Hsiong SX, et al. The effects of processing methods upon mechanical and biologic properties of porcine dermal extracellular matrix scaffolds. Biomaterials 2010;31(33): 8626–33. [39] Nagata S, Hanayama R, Kawane K. Autoimmunity and the clearance of dead cells. Cell 2010;140(5):619–30. [40] Zhang Q, Raoof M, Chen Y, Sumi Y, Sursal T, Junger W, et al. Circulating mitochondrial DAMPs cause inflammatory responses to injury. Nature 2010;464(7285):104–7. [41] Zheng MH, Chen J, Kirilak Y, Willers C, Xu J, Wood D. Porcine small intestine submucosa (SIS) is not an acellular collagenous matrix and contains porcine DNA: possible implications in human implantation. J Biomed Mater Res B Appl Biomater 2005;73(1):61–7. [42] Friedrich LH, Jungebluth P, Sjoqvist S, Lundin V, Haag JC, Lemon G, et al. Preservation of aortic root architecture and properties using a detergent– enzymatic perfusion protocol. Biomaterials 2013. [43] Paulin D, Li Z. Desmin: a major intermediate filament protein essential for the structural integrity and function of muscle. Exp Cell Res 2004;301(1):1–7. [44] Sierpinski P, Garrett J, Ma J, Apel P, Klorig D, Smith T, et al. The use of keratin biomaterials derived from human hair for the promotion of rapid regeneration of peripheral nerves. Biomaterials 2008;29(1):118–28. [45] Patel DM, Shah J, Srivastava AS. Therapeutic potential of mesenchymal stem cells in regenerative medicine. Stem Cells Int 2013;2013:496218. [46] Guo S, Dipietro LA. Factors affecting wound healing. J Dent Res 2010;89(3):219–29. [47] Freytes DO, Kang JW, Marcos-Campos I, Vunjak-Novakovic G. Macrophages modulate the viability and growth of human mesenchymal stem cells. J Cell Biochem 2013;114(1):220–9. [48] Bonenfant NR, Sokocevic D, Wagner DE, Borg ZD, Lathrop MJ, Lam YW, et al. The effects of storage and sterilization on de-cellularized and re-cellularized whole lung. Biomaterials 2013;34(13):3231–45. [49] Baiguera S, Del Gaudio C, Jaus MO, Polizzi L, Gonfiotti A, Comin CE, et al. Longterm changes to in vitro preserved bioengineered human trachea and their implications for decellularized tissues. Biomaterials 2012;33(14):3662–72. [50] Mulholland J, Aplin JD, Ayad S, Hong L, Glasser SR. Loss of collagen type VI from rat endometrial stroma during decidualization. Biol Reprod 1992;46(6):1136–43. [51] Scott JE, Thomlinson AM. The structure of interfibrillar proteoglycan bridges (‘shape modules’) in extracellular matrix of fibrous connective tissues and their stability in various chemical environments. J Anat 1998;192(Pt 3):391–405. [52] Scott JE, Stockwell RA. Cartilage elasticity resides in shape module decoran and aggrecan sumps of damping fluid: implications in osteoarthrosis. J Physiol 2006;574(Pt 3):643–50. [53] Brännström M, Diaz-Garcia C, Hanafy A, Olausson M, Tzakis A. Uterus transplantation: animal research and human possibilities. Fertil Steril 2012;97(6):1269–76.