Cell culture
Human osteoblasts (hOBs) were isolated from male patients undergoing hip replacement surgery (Queensland University of Technology Research Ethics approval number 1400001024) by explant culture48. The hOBs were cultured in expansion media consisting of minimum essential media alpha (MEM α), 10% fetal bovine serum (FBS) and 100 IU/ml penicillin and 100 µg/ml streptomycin (all Life Technologies, Mulgrave, VIC, Australia), before culture in osteogenic media consisting of expansion media supplemented with 50 µg/ml L-ascorbic acid, 10 mM β-glycerophosphate and 0.1 µM dexamethasone (all Sigma-Aldrich, Castle Hill, NSW, Australia). Patient-derived fibroblasts from radical prostatectomy cancer tissue (CAFs) were isolated and cultured in phenol red-free RPMI-1640 (Life Technologies) supplemented with 10% heat-inactivated FBS, 100 IU/ml penicillin, and 100 µg/ml streptomycin (Life Technologies), 1 nM testosterone (Sigma-Aldrich) and 10 ng/ml basic fibroblast growth factor (bFGF; Life Technologies)42. Human prostate-derived CD31 + lymphatic and blood vessel endothelial cells (MVECs) were isolated by enzymatic digestion with 0.25% collagenase II and 0.01% DNase (Worthington, Lakewood, NJ) of fresh radical prostatectomy tissue14. The BVECs were seeded into 10 µg/ml fibronectin (Sigma-Aldrich)-coated flasks and both the BVECs and HUVECs were cultured in the Endothelial Cell Growth Medium 2 Kit (EGM2; PromoCell, Heidelberg, Germany). MSCs were cultured in MEM α supplemented with 20% FBS and 100 IU/ml penicillin and 100 µg/ml streptomycin (Life Technologies)25. PC-3 cell lines (ATCC, MA, USA) were cultured in 2D and 3D to assess morphology and proliferation. PC-3-Luc1 was stably transduced to express luciferase using a pLenti6/V5-D-TOPO (Life Technologies) lentivirus4, and PC-3-Luc2 was obtained from collaborating scientists and routinely cultured in our laboratory. All PC-3 and LNCaP (ATCC) cells were maintained in phenol red-free RPMI-1640 (Life Technologies) supplemented with 10% FBS and 100 U/ml penicillin and 100 µg/ml streptomycin (Life Technologies), referred to below as ‘PCa media’, and used in passages 45, 32 and 14 for PC-3, PC-3-Luc1, and PC-3-Luc2, respectively. LNCaP-luc cells were used at passage 46. PC-3 cells were seeded at 3,000 cells/cm2 for 2D experiments and 350,000 cells/cm3 for 3D experiments. Cells were cultured for 7 days in 2D and 14 days in 3D. PC-3 and LNCaP cell line authentication was performed by the Genomics Research Center at QUT using short tandem repeat profiling with a ≥80% match for PC-3 (ATCC® CRL-1435™) and LNCaP (ATCC® CRL-1740™), respectively.
3D cell encapsulation
Cells were encapsulated in photocrosslinkable semi-synthetic gelatin-methacryloyl (GelMA)-derived hydrogels49. A cell suspension of 500,000 cells/ml was prepared in 4% w/v GelMA, containing 0.05% w/v of a water-soluble photo-initiator (Irgacure, IC2959; BASF, Germany). The precursor cell solution was transferred into a custom-made, rectangular Teflon mold, covered with a glass slide, and crosslinked using a CL-1000 UV cross-linker (UVP Upland, California, USA) with 365 nm wavelength tubes for 12 min (exposed intensity of 2.5 mW/cm2 on hydrogel surface). After polymerization, cell-laden hydrogels were removed from the mold, cut into 2 × 4 × 5 mm pieces, and transferred to 24-well plates for culture in PCa media. The media was changed every 3–4 days.
Microscopy
Microscopy images of GelMA hydrogels were recorded at day 1, 4, and 7 in 2D and day 1, 7, 14 in 3D using a phase-contrast IX73 microscope from Olympus at 10× and 20× magnifications.
DNA content
For cellular DNA content analysis, the cells were grown in 2D and the cell-laden 3D hydrogels were frozen at −80 °C for at least 48 h after two washes in phosphate buffer saline (PBS). Next, 350 µL of Proteinase K (Invitrogen, Australia) dissolved in phosphate-buffered EDTA (PBE) at 0.5 mg/mL was added to each sample and heated at 60 °C for 12 h using a block heater. The solution was diluted in PBE at a ratio of 1:4 for 2D samples and 1:2 for 3D samples and dispensed in duplicates (100 µL) into black 96-well plates (Corning, Australia). PicoGreen dsDNA quantitation (Invitrogen, Australia) working solution (100 µL) was added. After 5 min of incubation in the dark, the fluorescence (excitation 485 nm, emission 520 nm) was measured using a FLUOstar Omega plate reader (BMG LABTECH, Australia). A standard curve of known λ DNA concentrations ranging from 10 ng/mL to 1 µg/mL was used to calculate the final DNA content of the samples. Data are presented as mean ± SEM, n = 3–6.
Animal experiments
All animal experiments were approved by the Queensland University of Technology Animal Ethics Committee (approval number 130000025) in accordance with the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes. Male NOD-scid IL2Rγnull (NOD.Cg – Prkdcscid Il2rgtm1Wijl Hprtb-m3/EshJ; NSG) mice were obtained from the Translational Research Institute (TRI) in-house breeding colony at 4–6 weeks of age and acclimatized for up to 3 weeks. Animals were maintained under specific pathogen-free and temperature-controlled conditions. Sterilized food and water were provided ad libitum and mice were kept on a 12 h light−dark cycle.
Intraprostatic injection
A midline skin and peritoneal incision were made on the lower abdomen of male NSG mice. The urinary bladder and seminal vesicles were externalized. A cotton swab was used to further extend the bladder and seminal vesicles to reveal the prostate and perform the intraprostatic injection of the human prostate cells into the dorsal lobe of the mouse prostate. PC-3-Luc or LNCaP-Luc cells (250,000) were injected directly into the mouse prostate with or without accompanying CAFs (200,000) and human BVECs (50,000) in 50 µl of PBS50. The bladder and seminal vesicles were replaced and the incisions closed using sutures.
Bioluminescent imaging (BLI) analysis
Primary tumor formation and cancer cell metastasis were monitored weekly by in vivo BLI using a Xenogen IVIS Spectrum (PerkinElmer, Waltham, MA, USA). Images were acquired 15 min after intraperitoneal injection of 1.5 mg XenoLight D-luciferin potassium salt (PerkinElmer). At the experimental endpoint, the hTEBCs, murine prostate, bones, and organs were excised and analyzed using BLI within 20–30 min of D-Luciferin injection. Signal data were quantified using the Living Image v4.5.2 software (PerkinElmer) using the manual ROI tool to determine the amount of photons emitted for a given time. Signals above 50 counts for the ex vivo analysis were considered positive. The metastatic colonization of different tissues is indicated per mouse, except for the hTEBCs (two constructs per mouse) and kidneys, which are indicated per unit.
Confocal microscopy imaging
The in vitro tissue-engineered construct was stained with 2 µg/mL fluorescein diacetate (FDA) and 20 µg/mL propidium iodide (PI; both from Life Technologies) to detect living and dead cells, respectively. Additional constructs were fixed overnight at 4 °C in 4% paraformaldehyde (Sigma) then stained with 0.3 U/mL rhodamine 415-conjugated phalloidin and 2.5 µg/mL 4′6-diamidino-2-phenylindole (DAPI; Life Technologies). Samples were imaged with a Leica SP5 confocal scanning laser microscope49.
Scanning electron microscopy (SEM) imaging
For SEM imaging, the in vitro tissue-engineered constructs were fixed with 3% (v/v) glutaraldehyde in 0.1 M sodium cacodylate buffer (pH 7.3) at 4 °C until further processing could be performed. The fixed samples were dehydrated, critical point dried, coated with a gold sputter, and imaged with a Zeiss Sigma VP Field Emission SEM.
Generation of flat hTEBC
Flat scaffolds of medial grade polycaprolactone (mPCL) were fabricated via melt electrowriting5. Scaffolds with a pore size of 250 µm and comprised of 40 micrometer-sized mPCL layers were trimmed to pieces of 4 × 4 mm with a thickness of 400 µm. The scaffolds were coated with calcium phosphate (CaP) by immersion in highly saturated Simulated Body Fluid (SBF 10×) containing 1 M NaCl (ChemSupply, Gillman, SA, Australia), 5 mM KCl, 25 mM CaCl2.2H2O, 5 mM MgCl2.6H2O (all from Merck, Darmstadt, Germany) and 10 mM Na2HPO4 (Sigma-Aldrich), adjusted to pH 6 with NaHCO3 (Sigma-Aldrich) for 30 min at 37 °C. The SBF 10× incubation was repeated twice, replacing with fresh SBF 10× each time51. The mPCL-CaP scaffolds were seeded with 600,000 hOB/scaffold in 30 µl serum-free MEMα and cultured in well-plates for 2 weeks in expansion media followed by 11 weeks in osteogenic media. Throughout the course of the scaffold culture, hOBs migrated from the scaffold onto the tissue-culture plastic of the well plates and formed a hOB monolayer. The hOB monolayers were cultured together with the scaffold and were allowed to form a hOB cell sheet.
Star-shaped polyethylene glycol (sPEG)-heparin gels were prepared with a total of 100,000 cells (HUVECs:MSCs:hOBs at a ratio of 10:1:1) using gels at a crosslinking degree of γ1.5. The hydrogels were functionalized with a 2:1 molar ratio of RGD to heparin23 onto mPCL-CaP scaffolds. The “sandwich” scaffold was assembled with the sPEG-heparin gels cast on an unseeded mPCL-CaP scaffold sheet, layered with an additional sheet of mPCL-CaP scaffold either with or without pre-seeded and cultured hOBs.
hTEBC implantation procedure
Immediately prior to subcutaneous implantation, the mPCL-CaP in vitro constructs were combined using 60 µl fibrin glue (TISSEEL™ kit, Baxter Healthcare, Australia) embedded with 30 µg rhBMP-7 (Olympus Biotech Corporation, USA). Additionally, the rhBMP-7 and fibrin glue were combined with a hOB cell sheet isolated from the well plates in which the scaffolds were cultured for the hOB laden in vitro constructs. The hOB cell sheet was included as an additional source of hOBs in the hTEBC preparation. The sandwich scaffolds pre-seeded with hOBs were implanted into the left-back of the mouse whereas the unseeded scaffolds were implanted into the right-back of the mouse. To implant the scaffolds, two longitudinal incisions were made on the skin on the left and right back of the mouse. Subcutaneous pockets were created using blunt scissors to gently separate the subcutaneous space. The prepared hTEBCs were inserted into the prepared pockets and the incisions closed with wound closure autoclips (Kent Scientific Corporation, CT, USA). Autoclips were removed by 7−10 days when the surgical site had healed5.
Generation of tubular hTEBC
Tubular mPCL scaffolds with an internal diameter of 5 mm were fabricated via melt electrowriting and CaP-coated as described above5,7,25. The mPCL-CaP tubular scaffolds were seeded with 100,000 hOB/scaffold in 30 µl serum-free MEMα and cultured in well-plates for 2 weeks in expansion media followed by 7 weeks in osteogenic media. Throughout the course of the scaffold culture, hOBs migrated from the scaffold onto the tissue-culture plastic of the well plates and formed a hOB monolayer. The hOB monolayers were cultured together with the scaffold and were allowed to form a hOB cell sheet.
A prevascularized niche was generated using 4% GelMA-based hydrogels containing HUVECs and MSCs (10:1 ratio; 5.5 × 106 cells/ml total) 7 days before implantation and pre-cultured in EGM2 (PromoCell, Heidelberg, Germany) supplemented with 125 ng/ml SDF-1α, VEGF and FGF2 (Miltenyi Biotec, NSW, Australia). At the time of implantation, the vascular gel was placed inside the tubular hOB scaffold. The hOB cell sheet was mixed with 30 µl fibrin glue (TISSEEL™ kit, Baxter Healthcare, Australia) and 20 µl of rhBMP-2 (1.5 µg/µl; INFUSE®, Medtronic25. Two tubular hTEBC were implanted into left and right subcutaneous pockets on the back of male NSG mice as described above.
Micro CT
Ex vivo analysis was performed on the fixed hTEBCs using a high-resolution µCT scanner (µCT 40, Scanco Medical AG, Switzerland) and scanned at a voxel size of 16 µm. Samples were evaluated at a threshold of 220, a filter width of 0.8, and a filter support of 1. X-ray attenuation was correlated to the sample density with a standard curve generated after scanning hydroxyapatite phantoms with a known mineral density. Bone volume (BV), total volume (TV), and bone volume fraction (BV/TV) were calculated.
Histology and immunohistochemistry
After necropsy, samples were immediately fixed in 4% paraformaldehyde (Sigma-Aldrich) overnight and then transferred to 80% v/v ethanol until further analysis. Bone samples were decalcified for up to 5 weeks in 10% EDTA (pH 7.4) at 37 °C before all samples were subjected to routine processing and were embedded in paraffin wax. Serial sections were used for hematoxylin and eosin (H&E) staining and immunohistochemistry as outlined in Supplementary Table S6. Endogenous peroxidase activity was quenched with 3% hydrogen peroxide (Sigma-Aldrich) for 15 min and non-specific binding sites were blocked with 2% bovine serum albumin (BSA; Sigma-Aldrich). Primary antibodies were diluted in the blocking buffer. Immunoreactivity was detected using the EnVision + Dual Link System-HRP Rabbit/Mouse kit (Dako, Glostrup, Denmark) and was color developed with liquid diaminobenzidine chromagen (Dako). Sections were counterstained with Mayer’s Hematoxylin (Sigma) before dehydration and mounting. Human tissue and murine tissue were used as positive and negative controls respectively, for validating that antibodies reacted with human and not murine tissues5. Images were captured using a Leica SCN400 or a 3D Histech Panoramic high throughput slide scanner.
Statistics and reproducibility
Data were analyzed using GraphPad Prism v8.1.1 (GraphPad Software, La Jolla, California, USA) and SPSS statistics 23 (IBM Corporation, NSW, Australia). Normally distributed data were analyzed using an unpaired t-test, whereas data that were not normally distributed was analyzed using a Mann−Whitney test with a p-value < 0.05 accepted as significant. Data are represented as box plots depicting the median, first and third quartile, minimum and maximum, and are overlaid with individual data points. 2D and 3D cell culture data were analyzed using a one-way ANOVA with a Tukey post hoc test with a p-value < 0.05 accepted as significant.
Reporting summary
Further information on research design is available in the Nature Research Reporting Summary linked to this article.
