Primary human tissue
Primary pancreas tissue was obtained by the Cambridge Biorepository of Translational Medicine (CBTM) from deceased organ donors from whom multiple organs were being retrieved for transplantation. Pancreas samples were taken via two routes: from donors during the organ retrieval operation (in which organs other than the pancreas were taken for transplant) or from pancreata which were initially removed for organ transplantation but were subsequently declined and allocated for research. Tissue samples were placed in cold University of Wisconsin organ preservation solution prior to transportation to the laboratory.
Donor tissue was taken after obtaining written informed consent from the donor’s family for studies approved by the NRES Committee East of England, Cambridge South for the Department of Surgery, University of Cambridge, REC reference; 15/EE/0152 and the NRES Committee East of England - Cambridgeshire and Hertfordshire Research Ethics Committee for the Department of Surgery, University of Cambridge, REC reference; 16/EE/0227. Pancreas cancer tissue was obtained from patients undergoing pancreatic resection surgery who had given full written informed consent for studies approved by the NRES Cambridgeshire 2 Research Ethics Committee for Human Research Tissue Bank, Addenbrooke’s Hospital, REC reference; 11/EE/0011 and NRES Committee London - Westminster Research Ethics Committee for the Department of Surgery, University of Cambridge, REC reference; 15/LO/0753. Samples were taken by clinical histopathologists after gross examination of the resected tissue. Pancreatic islets were obtained from the Scottish National Blood Transfusion Service (SNBTS) Islet Isolation Center (NRES West Midlands- South Birmingham Research Ethics Committee, REC reference; 16/WM/0093).
Isolated primary pancreatic ducts for qRT-PCR analysis were collected via two methods: either by manual handpicking of ductal fragments following pancreas tissue digestion (as detailed below) or via surgical dissection of the main pancreatic duct from the pancreata allocated for research. Briefly, the common bile duct was separated from surrounding tissue and followed towards the ampulla of Vater where it connects to the primary pancreatic duct. The primary pancreatic duct was then separated from surrounding tissue and a segment was isolated.
Generation and culture of hPO
Handling and processing of samples was performed according to HTA guidelines. To generate organoid cultures, approximately 5 mg of pancreas sample was manually minced and further dissociated with the gentleMACS dissociator (Miltenyi Biotec) for a total of 2 min. Minced tissue was washed twice in Wash medium [Dulbecco’s Modified Eagle Medium (DMEM), high glucose, GlutaMAX, pyruvate supplemented (Life Technologies) with 1% Foetal Bovine Serum (FBS) (Life Technologies) and 1% Penicillin/Streptomycin (10,000 U/mL) (Life Technologies)] and digested in 40 mL of Digestion solution [Collagenase Type I (Sigma-Aldrich) and Dispase II (Life Technologies) at a concentration of 0.125 mg/mL in DMEM containing 0.1 mg/mL DNase I (Sigma-Aldrich)] and placed at 37 °C for 1 to 2 h. Isolated ducts were either hand-picked with a pipette or the whole digestion mixture was filtered with a 100 μm pore nylon cell strainer (Falcon). Ductal fragments were washed in Basal medium [Advanced DMEM/F12 (Life Technologies) supplemented with 1% Penicillin/Streptomycin, 1% Glutamax 100x (Life Technologies), and Hepes (Life Technologies) 10 mM] and spun at 200 g for 5 min. The cell pellet was mixed with reduced growth factor BME 2-RGF (Basement Membrane Extract Type 2 3533-010-02; AMSBIO, Cultrex), seeded in a 24 well plate and overlayed with the optimised hPO expansion medium (hPO-Opt.EM), unless specified otherwise. BME 2-RGF was used as ECM for all experiments except for those specified in Fig. 5 and Additional file 1: Figure S5 in which hPOs generated in BME 2-RGF were compared with those cultured in the chemically defined hydrogel. hPO-Opt.EM composition: [Basal medium (described above) supplemented with 1X N2 and 1X B27 (both from GIBCO), 1.25 mM N-Acetylcysteine (Sigma-Aldrich), 10% RSPO1 conditioned serum-free media (homemade as previously described ), 10 nM [Leu15]-Gastrin I human (Sigma-Aldrich), 50 ng/mL EGF (Peprotech), 25 ng/mL Noggin (Peprotech), 100 ng/mL FGF10 (Peprotech), 10 mM Nicotinamide (Sigma-Aldrich), 5 μM A83.01 (Tocris), 10 μM FSK (Tocris) and 3 μM PGE2 (Tocris)]. hPO-Opt.EM was supplemented with 10 μM Rho Kinase inhibitor (Y27632, Sigma-Aldrich) during the first 7 days. After 14 days, passaging was performed as previously described . Cryopreservation of established organoids was conducted as previously described .
Cryopreservation of pancreas tissue
Samples were manually minced and further dissociated using the gentleMACs dissociator (Miltenyi) for a total of 2 min. Samples were resuspended in 1 mL Recovery Freezing medium [(Dulbecco’s Modified Eagle Medium (High Glucose), fetal bovine serum, and DMSO (10%);Gibco)] and cryopreserved in a Cell freezing container at -80 °C. To initiate hPO generation after cryopreservation, the sample was thawed at 37 °C and washed twice in Wash medium. The procedure was then conducted as described above, beginning with the addition of Digestion solution.
Dispersion of hPOs to single cells
Preparation of single cell mixtures was performed as previously described ; briefly, confluent hPO wells were collected and washed with cold Basal medium. Cells were centrifuged at 200 g for 5 min and resuspended in 1 mL of pre-warmed trypsin (TrypLE™ Express Enzyme (1X)-Thermofisher). Organoids were pipetted using a narrowed Pasteur 10 times and incubated for 5 min at 37 °C to make single cells. After incubation, the cells were pipetted 10 times and checked for single cells. Digestion was stopped by adding cold Basal medium and the digest was filtered through a 40 μm pore nylon cell strainer (Falcon) to remove doublets.
Doubling time calculation
Doubling time was calculated as follows; the hPO cultures were dissociated into single cells as described above. Cell numbers were counted by trypan blue exclusion at the indicated time points. From the basic formula of the exponential curve y(t) = y0 x e (growth rate x t) (y = cell numbers at final time point; y0 = cell numbers at initial time point; t = time) the growth rate was derived. The doubling time was calculated as doubling time = ln (2)/growth rate for each time window analysed.
Lentiviral transduction and flow cytometry sorting
hPOs were expanded to passage 3, after which organoids were made into single cells as described above. 1 × 105 cells were resuspended in virus infection medium containing [CMV-GFP-T2A-Luciferase pre-packaged virus (Systems Bioscience, Cat. No. BLIV101VA-1) at a multiplicity of infection (MOI) of 5 (5x105u/μl) with 1:200 TransDux (System Bioscience) and 1:5 MAX Enhancer (Systems Bioscience) with hPO-Opt.EM]. The cell suspension was added to a 24 well plate, spun at 32 °C at 600 g for 10 min and then incubated at 37 °C for 6 h. Cells were then transferred to a 15 mL centrifuge tube, washed twice with Basal Medium and seeded in a 48 well plate with BME 2 and overlayed with hPO-Opt.EM supplemented with Rho Kinase inhibitor.
After 2 passages, cells were again subjected to a single cell dissociation as described above. Single cell preparations, along with negative controls (non-transduced hPOs) were sorted using a MoFlo cell sorter. GFP+ cells were seeded into 48 well plates with BME 2 and hPO-Opt.EM medium (with Rho Kinase inhibitor for the first 7 days). Organoids were expanded for 2 passages and imaged with the Evos Fl Imaging system (Thermofisher) for expression of GFP.
Generation of clonal cultures
hPO cultures were initiated from ductal fragments as described above, and allowed to grow for 10–14 days. P0 hPOs were made into single cells as described above. Single cells were centrifuged at 500 g for 5 min, re-suspended in BME 2 and then seeded in a 48 well plate at a density of 300–500 cells/well and allowed to expand for 15–20 days. Individual organoids were then picked out and reseeded (1 organoid per BME 2 drop). The single organoid was allowed to expand and then passaged as normal . For WGS 2–4 confluent wells (of a 24-well plate) were collected for each clone, snap frozen in PBS and submitted for genome sequencing.
Cells were lysed using a commercially available kit (Arcturus PicoPure DNA extraction kit; Thermo Fisher Scientific, ca. No. KIT0103). The samples were then sent for library construction and paired-end whole genome sequencing (150 bp) using Illumina XTEN® machines. Sequences were aligned to the human reference genome (NCBI build 37) using BWA-MEM. This resulted in a read depth of approximately ~35x per sample (see Additional file 3: Table S2).
Variant calling was performed using the CaVEMan algorithm . CaVEMan operates using a naive Bayesian classifier to derive the probability of all possible genotypes at each analysed nucleotide. For each sample CaVEMan was run using DNA sequenced from splenocytes from the same donor as a matched normal to identify germline SNPs. CaVEMan requires pre-input copy-number options, which were set to major copy number 5 and minor copy number 2 for normal clones, as we find this maximizes detection sensitivity. After variant calling we applied post-processing filters. We filtered against a panel of unmatched normal samples to remove single-nucleotide polymorphisms (SNPs) commonly present in the population. We also applied two filters designed to remove mapping artefacts associated with BWA-MEM: the median alignment score of reads supporting a mutation should be greater than or equal to 140, and below half of these reads should be clipped. Copy-number changes were called using the allele-specific copy number analysis of tumours (ASCAT) algorithm . The same matched normal sample was used as for calling single nucleotide variants with CaVEMan.
Culture of hPC
Human pancreas cancer organoids were cultured as previously described in Boj et al. . Briefly, the tumour sample was minced and placed in tumour digestion medium [Collagenase type II (5 mg/mL) made up in tumour organoid culture medium (hPC-EM)]. hPC-EM composition: [Basal medium (described above) supplemented with 1X N2 and 1X B27 (both from GIBCO), 1.25 mM N-Acetylcysteine (Sigma), 10 nM gastrin (Sigma), 50 ng/mL EGF (Peprotech), 40% Wnt3a conditioned medium (homemade), 10% RSPO1 conditioned media (homemade), 100 ng/mL FGF10 (Peprotech), 100 ng/mL Noggin (Peprotech), 10 mM Nicotinamide (Sigma), 0.5 μM A83.01 (Tocris), 1 μM FSK (Tocris) and 10 μM Rho Kinase inhibitor (Y27632, Sigma Aldrich)] and was digested overnight at 37 °C. The digest was spun at 300 g for 5 min and washed in Advanced DMEM/F12. The cell pellet was mixed with reduced growth factor BME 2, seeded in a 24 well plate and overlayed with hPC-EM medium. IPMN-derived tumour organoids were cultured in hPC-EM while PDAC-derived organoids were cultured in hPC-EM with 1 μM PGE2 (Tocris).
Chemically defined hydrogel culture
For the chemically defined dextran-based hydrogel (DEX-hydrogel) used in Fig. 5 and Additional file 1: Figure S5, SG-Dextran (Cellendes Cat. No. M91–3) and RGD Peptide (Cellendes Cat. No. 09-P-001) were used with a thiol-modified hyaluronic acid cross-linker. Thiol-modified hyaluronic acid was prepared as previously described , except that 4-(4,6-dimethoxy-1,3,5-triazin-2-yl)-4-methylmorpholinium chloride (DMTMM, TCI Chemicals) was used instead of N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (EDC) and N-hydroxysuccinimide (NHS) for attachment of cystamine to hyaluronic acid (Lifecore) with an average molecular weight of 57 kDa . Additionally, tris (2-carboxyethyl) phosphine (TCEP, Sigma-Aldrich, Cat. No. C4706) was used instead of dithiothreitol (DTT) to reduce the disulfide bond of the attached cystamines. Thiol-modified hyaluronic acid was purified by extensive dialysis against phosphate buffer at pH 3–5. This procedure yielded a modification of 12% of the D-glucuronic acid and N-acetyl-D-glucosamine disaccharides of hyaluronic acid with thiol groups as determined with the assay as previously described .
For preparation of 50 μl DEX-hydrogel; buffer, water and SG-Dextran of the SG-Dextran Kit (Cellendes Cat. No. M91–3) were used; 3 μl buffer (10x CB (pH 7.2)), 12.5 μl Water, 3.4 μl of SG-Dextran (30 mmol/L thiol-reactive groups) and 2.5 μl of RGD Peptide (20 mmol/L thiol groups) were combined and incubated for 20 min at room temperature. Thereafter, 20 μl organoid fragments were added and hydrogel formation was initiated by adding 8.6 μl of thiol-modified hyaluronic acid (50 nmol of thiol groups). The hydrogel/organoid suspension was seeded into 24 well plate during the 8 min pre-gel period and placed in the 37 °C incubator. 30 min after the initiation of crosslinking, the hydrogels were overlayed with the appropriate culture medium.
Passaging of organoids grown in DEX-hydrogel was achieved by first digesting the hydrogels with Dextranase (Cellendes Cat. No. D10–1) for 30–40 min at 37 °C according to the manufacturer’s recommendations. Once the gel was digested, organoids were fragmented by passing through a syringe with 27ga needle 3–5 times. Organoid fragments were washed 4 times with Basal medium and twice with Basal medium containing 11 mg/mL Dextran 6 (Carl Roth; Cat. No. 7615.1) to remove any Dextranase contamination. After the first passage, hPOs in DEX-hydrogel were cultured with hPO-Opt.EM medium supplemented with 10 mg/mL Dextran 6, acting as a competitive inhibitor to Dextranase, to inhibit gel degradation from leftover contaminating Dextranase; fresh medium was applied every day for 3 days post-passaging.
Chromosome counting of organoid cells was performed as previously described . Briefly, 24 h post-passaging, hPOs were incubated with 0.1 μg/mL KaryoMAX Colcemid solution in PBS (Gibco) for 24 h. hPOs were dissociated into single cells as described above, and were subsequently incubated in 1 mL of 0.075 M KCl (Fisher Chemicals) at 37 °C for 10 min. Cells were then fixed in a solution of 3:1 MeOH:Acetic Acid (VWR Chemicals) which was added dropwise while shaking. After fixation, the solution was dropped onto Superfrost Microscope Slides (VWR) for chromosomes to spread. The slide was then allowed to dry and mounted with Vectashield-Dapi (Vector Laboratories) and a coverslip.
RNA extraction of all material (organoids and primary tissue) was performed using an RNA extraction kit (Qiagen), and as previously described . A complete list of the primers used can be found in Additional file 5: Table S4. All qRT-PCR data is displayed as mean + SEM, with each data point representing a separate donor line. Values are given relative to the expression of the housekeeping gene Hypoxanthine-guanine phosphoribosyltransferase (HPRT).
Organoid and tissue fixation and paraffin embedding
Organoids were collected from wells, washed with Basal medium and fixed in 10% neutral-buffered formalin (Sigma-Aldrich) for 30–40 min on ice. Primary tissue and xenograft-derived tissue was placed directly into 10% formalin and fixed overnight at room temperature. Organoids and tissue were then embedded in paraffin as follows: samples were dehydrated through a series of graded-ethanol solutions, followed by xylene (Fisher) and finally embedded in paraffin. Sections were cut at 5 μm thickness and were placed at 60 °C for 2–24 h.
H&E of organoids, primary tissue and xenograft tissue
Paraffin slides were rehydrated with xylene, and then decreasing ethanol concentrations (100–50%) and water. Slides were then immersed in Haematoxylin (Sigma-Aldrich), washed and dehydrated in increasing ethanol concentrations (50–100%), a 10 s wash step of Eosin (Sigma) and finally xylene. The slides were then mounted with DPX mounting solution (Fisher).
Immunostaining of organoids and tissue sections
Organoids were washed in PBS with 0.05% BSA following formalin fixation as described above. Tissue sections were rehydrated as described above for H&E staining. Following rehydration, slides were washed with PBS and subjected to antigen retrieval by heating to 80 °C in 10 mM Sodium Citrate (Sigma), pH 6 for 20 min. Organoids and sections were incubated with blocking solution [Triton X100 (1% for nuclear antibodies and 0.1% for membrane and cytoplasmic antibodies), 1% BSA, 2% donkey serum] for 2 h at room temperature. Primary antibodies were applied at specified dilutions overnight at 4 °C. Organoids and tissues were washed with PBS, and appropriate secondary antibodies were applied for 2 h at room temperature, washed with PBS and nuclei were counterstained with Hoechst33342 (Molecular Probes, Life Technologies). Please refer to Additional file 5: Table S4 for primary and secondary antibody information and dilutions.
Brightfield and confocal imaging
Brightfield imaging of organoids was performed using a Leica M80 stereo microscope (Leica Microsystems) and a Leica DMIL LED microscope (Leica Microsystems). H&E images were taken using a Leica DM400B LED microscope (Leica Microsystems). Karyotypes and IF staining were imaged using a SP8 White Light inverted confocal microscope (Leica Microsystems) or with a Leica DMI3000 fluorescent inverted microscope (Leica Microsystems). Optical sections were acquired at 3 μm intervals. Images were acquired with Leica application suite X Software and processed using Fiji.
Mouse xenograft studies
All animal experiments were performed in accordance with UK Home Office regulations (UK Home Office Project License number PPL 70/8702 and PPL P57643EBB). Immunodeficient NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ (NSG) mice which lack B, T and NK lymphocytes [51, 52] were bred in-house with food and water available ad libitum pre- and post-procedures. Male and female animals were used, aged approximately 6–8 weeks (average weight 20 g/each). Animals were allocated at random to experimental groups, tissue sections obtained from animals were processed, stained and analysed without reference to the identity of the animal groups.
hPO and hPC-org cultures were expanded in order to inject 5 × 105-1 × 106 cells per mouse. Organoids were mechanically dissociated as described for normal passaging and resuspended in the appropriate injection medium as outlined in Additional file 4: Table S3. The cells were loaded into a 250 μL glass gastight syringe (Hamilton) with removable 26ga blunt needles (ESSLAB) for injection into the kidney capsule or custom made 26ga point needles, bevelled at a 60o angle (ESSLAB) for injection into the pancreas capsule. Mice were anesthetised using isoflurane gas and the left side of the abdomen or peritoneal abdomen was shaved and cleaned with disinfectant. During the procedure, the mice were kept under anaesthesia and were kept on heat pads at 37 °C. Injections were performed as described below into the kidney capsule or pancreas capsule. Following the xenograft procedure, all animals were kept alive for either 1 or 3 months after which they were humanely euthanised under terminal anaesthesia by inhalation of isoflurane. Tissue was then retrieved for further histological analysis.
Kidney capsule injections
An incision of the skin was made near the anatomical position of the kidney, the kidney was localised and a further incision of the abdominal wall was made to expose the kidney. The kidney was gently pushed out of the abdomen and kept wet with sterile saline. A small incision to the kidney capsule was made with a sharp needle, then 20 μL of the organoid suspension was injected under the capsule using the blunt needle syringe. A sterile cotton bud was used to apply pressure to the point of insertion to stop bleeding and prevent cell leakage. The kidney was then gently placed underneath the muscle wall. The muscle wall was sutured first using continuous suturing with 5–0 vicryl sutures and interrupted sutures were used to close the skin layer afterwhich 9 mm autoclip wound clips (Harvard Apparatus) were placed on the skin to keep the sutures intact.
Pancreas capsule injections
An incision of the skin and abdominal wall was made along the midline of the abdomen to expose the visceral organs. The pancreas was exposed and kept wet with sterile saline. Using a sterile cotton bud for traction, organoids were injected into the tail of the pancreas, through the parenchyma, and placed between the inter-lobular space. The cotton bud was then used to stop leakage by applying pressure for 10–15 s. The pancreas was gently placed back to the correct anatomical position and the abdomen wall and skin were sutured using continuous suturing with 5–0 vicryl sutures.
All summary data are presented as mean ± SEM. Statistical tests were performed using Graphpad Prism software (GraphPad 8.1). Sample size (n) values used for statistical analyses are provided in the relevant figure legends. Student’s two-tailed unpaired t-test (or paired where specified) was performed to assess differences between two groups. When performing a t-test, we assumed normality and equal distribution of variance between groups. Significance was set at P < 0.05 for all experiments.