Georgakopoulos, Nikitas, Prior, Nicole, Angres, Brigitte, Mastrogiovanni, Gianmarco, Cagan, Alex, Harrison, Daisy, Hindley, Christopher, Arnes-Benito, Robert, Siong-Seng Liau, Curd, Abbie, Ivory, Natasha, Simons, Benjamin, Inigo Martincorena, Wurst, Helmut, Saeb-Parsy, Kourosh, and Meritxell Huch
Additional file 1: Figure S1. Optimisation of hPO-Opt.EM culture medium and its expansion potential compared with published pancreatic organoid culture systems. A) To obtain an optimised medium to support hPO isolation and growth, incremental changes were made to previously published protocols. Expansion graphs show the time hPOs survived in vitro for n = 4 donors in mouse pancreatic organoid (mPO) medium [20] supplemented with TGFb inhibitor (top graphs) or in our previously reported medium supplemented with TGFb inhibitor, PGE2 and Wnt-conditioned medium containing 10% serum [23] (middle graphs) or in this new optimised, chemically-defined medium (hPO-Opt.EM) containing TGFb inhibitor, PGE2, FSK and no-FBS (bottom graphs). The ability of the hPOs to expand is indicated by passaging events (cirlces), arrows indicate ongoing cultures, capped lines indicate cultures that deteriorated. While the first two conditions were not able to sustain long-term culture, the hPO-Opt.EM medium demonstrates a much greater expansion potential. B-D) Comparison of the hPO-Opt.EM medium to the chemically-defined hPO medium published by Loomans and colleagues during the course of this project [22]. Although both media enable the initial generation of hPOs, the hPO-Opt.EM medium supports long-term culture to a much greater extent than Medium [22]. B-C) Representative images of hPO cultures derived from fresh pancreas tissue using the the hPO-Opt.EM medium or Medium [22] at B) passage 0 (P0, 8-days post derivation; magnification in lower panels; n = 3), or C) at Passage 3 (top) and Passage 4 (bottom) in two independent donors. D) Graph shows the expansion potential of hPOs cultured with the hPO-Opt.EM medium or the medium published by Loomans and colleagues [22] (arrows indicate ongoing cultures, capped lines indicate cultures that deteriorated). Figure S2. Human pancreas organoids (hPOs) can be derived from fresh and cryopreserved pancreas tissue and are amenable for genetic manipulation. A) Organoid cultures derived from fresh tissue (left) or tissue cryopreserved at the time of collection (right). The organoid formation efficiency from cryopreserved tissue was reduced, yet in all cases, cultures exhibited similar expansion rate to hPOs derived from fresh tissue. Experiments were performed in n = 3 independent donors, with similar outcomes. Representative images are shown. B) Organoid formation from fresh tissue is more efficient than from cryopreserved tissue, the number of organoids formed following the isolation of ductal fragments from either fresh tissue (253 ± 58 organoids; black circles) or cryopreserved tissue (25 ± 3 organoids; blue squares) is shown. Ductal fragments were seeded in a 50 μl BME 2 drop and quantified at P0. Data presented as mean ± SEM. C) hPOs derived from fresh or cryopreserved tissue expand at similar rates (circle = passage). D) Following mechanical dissociation all organoid fragments are capable of forming a new organoid (passage). Representative images of hPO culture during passaging are shown. E) The hPO culture system supports expansion from dissociated single cell suspensions, hPO cultures derived from single cells exhibit similar colony formation efficiency at early as well as late passages (n = 4 independent donors). F-G) Genetic manipulation of hPOs. hPOs were dissociated to single cells at passage 3 and transduced with a lentiviral vector carrying a GFP reporter gene. Following viral transduction the single cells formed hPOs and were expanded for 2 passages. F) hPOs were dispersed into single cells again and FACS sorting was used to select for GFP-positivity. The GFP+ cells were isolated and expanded as genetically modified hPOs for a further 2 passages. G) Representative images of genetically modified organoids at passage 8 are shown (n = 2). Figure S3. hPOs derived from fresh and cryopreserved samples expand as a single cell-layer epithelium of ductal cells and are phenotypically indistinguishable. Comparison of hPOs derived from A) fresh tissue (collected at P3) or B) cryopreserved tissue (collected at P2) from the same donor. Brightfield images (upper panels) indicate that hPO cultures expand efficiently as cystic structures regardless of whether the original tissue is fresh (A) or cryopreserved (B). H&E stainings reveal that hPOs derived from cryopreserved tissue maintain the single cell-layer epithelial architecture seen in the original donor tissue and in hPOs derived from fresh tissue (middle panels). Immunofluorescence stainings demonstrate that hPO cultures maintain ductal identity (KRT19 and SOX9), cellular polarisation (F-Actin) and express PDX1 (lower panels) regardless of the original tissue being fresh (left) or cryopreserved (right). Figure S4. Optimisation of hPO transplantation. A-B) hPOs were transplanted into NSG mice at different sites using a combination of different vehicles and ECMs, tissues were retrieved 1 and 3 months after injections. A) Table outlining engraftment conditions and engrafment success rates of hPO xenografts conducted with a combination of injection medium compositions and injection sites in order to achieve long-term survival of hPOs in vivo (NT – not tested). B) Representative H&E images of engrafted hPO cells at 1 month (upper panels) or 3 months (lower panels) show engrafted cells form ductal-like structures in vivo (G-graft, PN-pancreas). C) Xenografts of hPOs into the pancreas collected after 3 months show the engrafted cells retain SOX9 protein expression. Figure S5 hPOs cultured in optimised hPO-Opt.EM medium grow in a chemically defined hydrogel. A-B) Comparison of hPO cultures in the chemically defined dextran-based hydrogel (DEX-hydrogel) initiated from freshly isolated ducts overlayed with either A) hPO-Opt.EM medium or B) the medium published by Loomans and colleagues [22]. Both media support the formation of cystic structures, however, organoids in Medium [22] quickly deteriorated and could not be passaged. In contrast, organoids formed with the optimised hPO-EM medium could undergo expansion up to P4, after which they deteriorated. C) hPOs generated in hPO-Opt.EM medium expand more slowly in DEX-hydrogel as shown by the longer time taken for the cultures to reach confluency in order to be passaged than hPOs embedded in BME 2 (circles-passage events; arrows indicate ongoing cultures, capped lines indicate cultures that have deteriorated).