3D imaging of colorectal cancer organoids identifies responses to Tankyrase inhibitors

TL;DR: It is demonstrated that morphometric analyses can capture subtle alterations in organoid responses to Wnt inhibitors that are consistent with activity against a cancer stem cell subpopulation, highlighting the value of phenotypic readouts as a quantitative method to asses drug-induced effects in a relevant preclinical model.

Abstract: Aberrant activation of the Wnt signalling pathway is required for tumour initiation and survival in the majority of colorectal cancers. The development of inhibitors of Wnt signalling has been the focus of multiple drug discovery programs targeting colorectal cancer and other malignancies associated with aberrant pathway activation. However, progression of new clinical entities targeting the Wnt pathway has been slow. One challenge lies with the limited predictive power of 2D cancer cell lines because they fail to fully recapitulate intratumoural phenotypic heterogeneity. In particular, the relationship between 2D cancer cell biology and cancer stem cell function is poorly understood. By contrast, 3D tumour organoids provide a platform in which complex cell-cell interactions can be studied. However, complex 3D models provide a challenging platform for the quantitative analysis of drug responses of therapies that have differential effects on tumour cell subpopulations. Here, we generated tumour organoids from colorectal cancer patients and tested their responses to inhibitors of Tankyrase (TNKSi) which are known to modulate Wnt signalling. Using compounds with 3 orders of magnitude difference in cellular mechanistic potency together with image-based assays, we demonstrate that morphometric analyses can capture subtle alterations in organoid responses to Wnt inhibitors that are consistent with activity against a cancer stem cell subpopulation. Overall our study highlights the value of phenotypic readouts as a quantitative method to asses drug-induced effects in a relevant preclinical model.

Summary

Introduction

• Wnt/β-catenin signalling, as a result of activating mutations within the pathway, has a prominent role in the initiation and progression of colorectal cancer (CRC) [1, 2].
• In the absence of a Wnt ligand the multi-protein β-catenin destruction complex, formed of AXIN1/2, Adenomatous polyposis coli (APC) and glycogen synthase kinase (GSK3β), mark β-catenin for degradation [3].
• The biological and phenotypic complexity of 3D primary organoid cultures allow, in principle, the assessment of compound activity against a subset of inter-cellular signalling that is central to tumour growth.
• The authors generated CRC patient-derived organoids and studied their responses to Tankyrase inhibitors .

Materials

• Corning Growth factor-reduced Matrigel was purchased from VWR (#734–1101).
• Cell culture media were purchased from Invitrogen Life Technologies, and cell culture plastics from Nunc unless otherwise stated.
• All TNKSi compounds were supplied and synthesized by Merck Healthcare KGaA (Darmstadt, Germany).
• All stock solutions of compounds were reconstituted in DMSO.

Human tissue

• Surgically resected patient materials were obtained from University Hospital of Wales by the Wales Cancer Bank with written informed ethical consent from male and female patients (>16 years of age, above the UK age of valid consent) with known or suspected malignant disease and anonymised (WCB project reference #12/001).
• The Wales Cancer Bank has ethics approval as a Research Tissue Bank from the Wales Research Ethics Committee 3 (reference 16/WA/0256), and is licensed by the Human Tissue Authority under the UK Human Tissue Act (2004) to store human tissue, taken from the living, for research (licence 12107) [25].
• These approvals cover the collection of samples (including written consent), processing and storing samples across multiple collection and storage sites.
• All patient derived material was handled in concordance with HTA regulations.
• Histological sections of patient tissue were imaged at University Hospital Wales and remained anonymised.

Organoid culture

• The isolation of tumour organoids from patient material was processed as previously described by Sato et al., [20] with some refinements outlined below.
• Once digested, tissue was triturated in PBS at room temperature to release cell fragments from tissue.
• Cell fragments within the supernatant were then centrifuged for 5 min at 100 rcf, at 4˚C.
• Following Matrigel polymerisation, cells were overlaid with 500 μL of either “7+” or “Full” media and replenished every 4 days.

DNA extraction

• DNA extraction from organoid cultures was carried out using a QIAamp DNA Mini Kit following the manufacturer’s instructions.
• Patient blood samples were processed by the Wales Cancer Bank (WCB), genomic DNA was extracted from 4 ml of whole blood using the ChemagicSTAR automated cell lysis and DNA extraction workstation (Hamilton Company) housed within the All Wales Medical Genetics Service laboratory.
• The steps included tagmentation of gDNA, clean-up of the tagmented DNA, amplification, clean-up of the amplified DNA, hybridisation of probes, capture of the hybridized probes, second hybridization of probes, second capture, clean-up of the captured library, amplification of enriched library, clean-up of the enriched library and finally validation of the complete library.
• Following validation, the libraries were normalized to 10 nM, pooled together and clustered on the cBot™2 following the manufacturer’s recommendations.
• A custom Perl script was used to extract genes of interest from annotated variant files.

Western blot detection

• Total protein was extracted from organoids by addition of 500 μL of lysis buffer to whole organoids (0.02 M Tris-HCl, 2 mM EDTA, 0.5% v/v NP-40 (IPEGAL) in ddH2O containing 1x PhosSTOP phosphatase inhibitor and 1x complete Protease Inhibitor Cocktail ).
• The lysates were centrifuged at 8000 rcf for 15 min and proteins harvested in the supernatants.
• Samples were resolved on Novex NuPAGE 4–12% Bis-Tris PAGE gels, and then blotted onto nitrocellulose membranes using the Invitrogen iBlot Dry Blotting cassette system.
• After washing the blots were incubated with HRP-conjugated secondary antibodies for 1 h at room temperature.
• Bands were visualized using the enhanced chemoluminescence (SuperSignal West Dura; Pierce).

Organoid viability measurements

• Organoids in culture were gently dissociated to near-single cell populations using TrypLE (Life Technologies) before resuspension within growth factor-reduced Matrigel, and dispensed into white clear bottomed 96 well plates in 9 μL Matrigel per well (400 cells/ μL of Matrigel).
• Upon polymerisation of Matrigel, growth media containing a titration range of individual compounds or DMSO controls were added to each well prior to incubation at 37˚C.
• Masks generated from both channels facilitated the visualisation of main organoid structure, internal morphometries such as individual lumens, as well as nuclei per organoid.
• Multiple features extracted included counts, total area, solidity, and branching of nuclei, lumen and organoid masks.
• Results are shown as means ± standard deviations unless otherwise stated.

Quantitative real-time PCR

• Total RNA was extracted from organoids resuspended in trizol solution (Life Technologies) containing 125 μg/ml glycogen.
• CDNA was yielded from purified mRNA using ImProm II Reverse Transcription kit .
• Quantitative-RT PCR was then performed using the SensiFAST SYBR Green Hi-ROX master mix .
• Primers were designed (Sigma Aldrich) and listed in S2 Table.
• All samples were measured in triplicate, with gene expression normalised to GAPDH housekeeping gene.

Wholemount immunofluorescent staining

• Secondary antibodies were then added overnight at 4˚C, prior to counterstaining with Hoechst.
• In vivo organoid engraftment studies (also termed “organoid derived xenografts” (ODX)) were conducted using immune-deficient NOD/SCID gamma irradiated mice.
• Palpable tumours (> 5 mm) were counted for use in Kaplan-Meier analysis.
• Cardiff University’s Animal Welfare and Ethical Review Body reviewed the plan for animal studies prior to the grant of the Home Office License (30/3279).

Results

• Establishment of tumour organoids from CRC patient material Surgically-resected CRC material was isolated from patients under informed consent.
• The organoid medium combinations shown in Fig 1A were used for subsequent studies.
• EC50 values in CRC lines ranged from 2nM to 1μM, but accurate values were hard to establish due to assay variance and also because many cells survived, even in the presence of high concentrations of compound, leading to reduced assay windows.
• Taken together, these results suggest that the pharmacological inhibition of TNKS results in an overall reduction of Wnt/β-catenin signalling in both TNKSi-sensitive and resistant organoid lines and that phenotypic responses could not be simply predicted based on the expression of biomarkers.

Discussion

• 3D organoids represent an attractive platform for the capture of the effects of therapies targeting signalling pathways involved in the regulation of complex morphometric interactions that are not well represented in clonal 2D culture.
• Numerous morphometric features, such as the distance of the matrix to central lumens and the number of nuclei per structure were quantified using automated algorithms.
• The identification of a set of morphometric parameters that accurately distinguish between compounds of different potency allowed putative biomarkers of TNKSi sensitivity to be compared with organoid phenotypic responses.
• A simple correlation between Wnt-dependence in culture and TNKSi sensitivity was not observed since 5 out of the 8 lines examined showed morphometric responses that tracked the cellular potency of the TNKS inhibitors (Fig 3B).
• August 18, 2020 15 / 20 suggest that modulation of Wnt signalling in a sensitive line was sufficient to limit the stemlike signature of organoid cell populations.

S1 Checklist.

• (DOCX) S1 Fig. Hematoxylin & Eosin staining.
• Novex Sharp pre-stained protein standards (Invitrogen; LC5800) were loaded as per manufacturers instructions.
• Projections of the Hoechst (Blue) and Phalloidin-rhodamine (Red) signal are overlaid with the cell and lumen mask .
• A Representative confocal images of organoids stained with an Lgr5 and Cytokeratin 20 antibody following six days of exposure to C1 (15 nM) or control (DMSO, 0.1%).

S2 Table. Primers used for qRT-PCR by SYBR green.

• Whole exome sequencing analysis for each organoid line.
• The mutations reported in the table are those found to have been previously described as likely oncogenic and curated within the COSMIC and cBioportal databases.

Acknowledgments

• The authors acknowledge all members of the University Hospital of Wales (UHW) Colorectal surgical team and UHW Histopathology team.
• The authors would like to thank the Wales Cancer Bank for providing access to all patient material and to patients for providing their consent to this study.
• The authors acknowledge their colleagues at the Wales Gene Park for their insight and expertise that assisted this research, and their technical and bioinformatic support in generating the NGS data.
• The authors acknowledge the support of Cellesce Ltd. The authors thank Victoria Marsh Durban and Anika Offergeld for manuscript proof reading and corrections.
• T.C.D accepts responsibility for the integrity and validity of the data collected and analysed.

Author Contributions

• Luned M. Badder, Andrew J. Hollins, Alan R. Clarke, Dirk Wienke, Leo S. Price, Paul H. Shaw, Trevor C. Dale, also known as Conceptualization.
• Luned M. Badder, Bram Herpers, Kuan Yan, Marc Naven, Kevin E. Ashelford, also known as Data curation.
• Alan R. Clarke, Paul H. Shaw, Trevor C. Dale, also known as Funding acquisition.
• Luned M. Badder, Andrew J. Hollins, Bram Herpers, Kuan Yan, Kenneth B. Ewan, Mairian Thomas, Jennifer R. Shone, Delyth A. Badder, Christina Esdar, Hans-Peter Buchstaller, also known as Investigation.
• Luned M. Badder, also known as Writing – original draft.

Figures

Fig 3. 3D image analysis demonstrates target affinity of tankyrase inhibitors in organoids. (A) Flow diagram of high-content multi-parametric image screen process. (B) Dose response curves acquired from high content imaging screen. Organoids were seeded as single cells in 384 well plates and exposed to Tankyrase inhibitors C1, C2, C3 for six days. Individual morphological measurements were generated from DAPI and F-actin channels. Principal component analysis was used to select the top 10 most discriminating features to separate compound-induced phenotypes between high and low doses of Tankyrase inhibitors. Selected parameters were used for feature space training, whereby Euclidean distance metrics were calculated between high and low doses of each compound for each organoid line. Euclidean distances were used as an indicator of similarity between low and high doses, with a low distance indicative of high phenotypic similarity between conditions. Data are presented as mean Euclidean distance from 8 replicate wells ± S.D.

Fig 2. Assessing the effects of Tankyrase inhibition on overall organoid viability. (A) Structures of three tankyrase inhibitors: MSC2524070A (C1), MSC2572070A (C2) and MSC2501490A (C3), with corresponding EC50 values. (B) Western blot of eight organoid lines treated with C1 (15 nM) and harvested

Fig 5. TNKSi-treated organoids enhance survival in an organoid-derived xenograft (ODX) mouse model. (A) Schematic diagram depicting organoid transplantation into mice. (B) A cohort of 16 NOD/SCID/ γ-irradiated mice were injected with pre-treated (30 nM C2 for 24 hours) Iso 75 organoids, or control (0.1% DMSO, 24 hours) organoids within one injection site per mouse. TNKSI-treated organoids showed a significant delayed formation of tumours compared to control conditions (Log-Rank (Mantel-Cox) test P = 0.032.

Fig 4. Tankyrase inhibition alters the proportion of stemness in CRC organoids. (A) Representative confocal images of organoids stained with a β-catenin antibody following six days of exposure to C2 or control (DMSO, 0.1%) showing a reduction in the intensity and numbers of cells showing β-catenin positivity following treatment (scale bar = 50 μm). (B). Quantitative real-time PCR analysis of relative mRNA expression of intestinal epithelial markers from organoids (Iso 50, Iso 72, Iso 75) cultured with tankyrase inhibitor C1 (15 nM, 6 days) or DMSO (0.1%). Markers associated with intestinal stem cell activity (LGR5, ASCL2), differentiation (KRT20, DKK1) alongside other members of the Wnt– signalling pathway were determined, with mRNA expression normalized to control conditions. Experiments were performed in biological repeats (n = 3) with error bars indicative of S.E.M (���, p�0.001; ��, p� 0.01; �, p�0.05, Paired Student’s t-test with Bonferroni correction).

Fig 1. Generation and establishment of tumour organoids from CRC patient material. (A) Summary of parental tumour sites and Dukes’ Stages, organoid media conditions, and mutation status of organoid lines identified in Whole Exome Sequencing analysis. (B) Representative Hematoxylin+Eosin (H&E) staining of FFPE tissue sections from primary colorectal tumour patient material and organoid counterparts. Organoids were cultured for a minimum of 6 days prior to fixation and embedding. Qualitative morphological similarities were observed between tumour and organoid pairs. Scale bar = 100 μm.

Table 1. Summary of EC50 values obtained by morphometric analysis.

Full text

RESEARCH ARTICLE
3D imaging of colorectal cancer organoids
identifies responses to Tankyrase inhibitors
Luned M. Badder
1,2☯
, Andrew J. Hollins
1,2☯
, Bram Herpers
3
, Kuan Yan
ID
3
, Kenneth
B. Ewan
1
, Mairian Thomas
4
, Jennifer R. Shone
1
, Delyth A. Badder
5
, Marc Naven
6
, Kevin
E. Ashelford
6
, Rachel Hargest
7,8
, Alan R. Clarke
2†
, Christina Esdar
9
, Hans-
Peter Buchstaller
9
, J. Mark Treherne
4
, Sylvia Boj
10
, Bahar Ramezanpour
10
, Dirk Wienke
9
,
Leo S. Price
3
, Paul H. Shaw
11
, Trevor C. Dale
ID
2
*
1 Cardiff University School of Biosciences, Cardiff, Wales, United Kingdom, 2 European Cancer Stem Cell
Research Institute (ECSCRI), Cardiff University, Cardiff, Wales, United Kingdom, 3 OcellO B.V., Leiden, The
Netherlands, 4 Cellesce Ltd, Cardiff Medicentre, Heath Park, Cardiff, United Kingdom, 5 Cellular Pathology
Department, University Hospital for Wales, Cardiff, United Kingdom, 6 Division of Cancer and Genetics,
School of Medicine, Cardiff University, Cardiff, United Kingdom, 7 Department of Colorectal Surgery,
University Hospital of Wales, Cardiff, United Kingdom, 8 Division of Cancer and Genetics, CCMRC, Henry
Wellcome Building, Cardiff University, Cardiff, United Kingdom, 9 Biopharma, Merck Healthcare KGaA,
Research & Development, Darmstadt, Germany, 10 Hubrecht Organoid Technology, Utrecht, The
Netherlands, 11 Velindre Cancer Centre, Cardiff, Wales, United Kingdom
☯ These authors contributed equally to this work.
† Deceased.
* DaleTC@cardiff.ac.uk
Abstract
Aberrant activation of the Wnt signalling pathway is required for tumour initiation and sur-
vival in the majority of colorectal cancers. The development of inhibitors of Wnt signalling
has been the focus of multiple drug discovery programs targeting colorectal cancer and
other malignancies associated with aberrant pathway activation. However, progression of
new clinical entities targeting the Wnt pathway has been slow. One challenge lies with the
limited predictive power of 2D cancer cell lines because they fail to fully recapitulate intratu-
moural phenotypic heterogeneity. In particular, the relationship between 2D cancer cell biol-
ogy and cancer stem cell function is poorly understood. By contrast, 3D tumour organoids
provide a platform in which complex cell-cell interactions can be studied. However, complex
3D models provide a challenging platform for the quantitative analysis of drug responses of
therapies that have differential effects on tumour cell subpopulations. Here, we generated
tumour organoids from colorectal cancer patients and tested their responses to inhibitors of
Tankyrase (TNKSi) which are known to modulate Wnt signalling. Using compounds with 3
orders of magnitude difference in cellular mechanistic potency together with image-based
assays, we demonstrate that morphometric analyses can capture subtle alterations in orga-
noid responses to Wnt inhibitors that are consistent with activity against a cancer stem cell
subpopulation. Overall our study highlights the value of phenotypic readouts as a quantita-
tive method to asses drug-induced effects in a relevant preclinical model.
PLOS ONE
PLOS ONE | https://doi.org/10.1371/journal.pone.0235319 August 18, 2020 1 / 20
a1111111111
a1111111111
a1111111111
a1111111111
a1111111111
OPEN ACCESS
Citation: Badder LM, Hollins AJ, Herpers B, Yan K,
Ewan KB, Thomas M, et al. (2020) 3D imaging of
colorectal cancer organoids identifies responses to
Tankyrase inhibitors. PLoS ONE 15(8): e0235319.
https://doi.org/10.1371/journal.pone.0235319
Editor: Ning Wei, University of Pittsburgh, UNITED
STATES
Received: November 1, 2019
Accepted: June 12, 2020
Published: August 18, 2020
Copyright: © 2020 Badder et al. This is an open
access article distributed under the terms of the
Creative Commons Attribution License, which
permits unrestricted use, distribution, and
reproduction in any medium, provided the original
author and source are credited.
Data Availability Statement: Components within
the manuscript (denoted as S1 Table within the
text) are held in a public repository https://www.
ebi.ac.uk/ena/data/view/PRJEB37607 where
sequencing information can be found. Other data is
all contained within the paper and/or Supporting
Information files.
Funding: This work was supported by Cancer
Research Wales PhD studentship (L.M.B), the
Cancer Research UK Cardiff Experimental Cancer
Medicine Centres (ECMC) (A.J.H). and Cancer
Research UK Programme Grant C1295/A15937H

Introduction
Aberrant canonical Wnt/β-catenin signalling, as a result of activating mutations within the
pathway, has a prominent role in the initiation and progression of colorectal cancer (CRC) [1,
2]. A number of feedback loops control β-catenin turnover and Wnt activation. In the absence
of a Wnt ligand the multi-protein β-catenin destruction complex, formed of AXIN1/2, Adeno-
matous polyposis coli (APC) and glycogen synthase kinase (GSK3β), mark β-catenin for deg-
radation [3]. As a result, the accumulation and subsequent translocation of β-catenin to the
nucleus is inhibited, preventing the downstream activation of target genes [4]. Components of
the destruction complex are also tightly regulated. AXIN1 and AXIN2 are concentration-limit-
ing components of the destruction complex. Levels of AXIN1/2 are post-transcriptionally reg-
ulated by tankyrases (TNKS1 and TNKS2), members of the poly (ADP-ribose) polymerases
(PARP) family of enzymes, which enhance Wnt signalling by targeting AXIN1/2 for degrada-
tion [5].
The inhibition of Wnt signalling has been validated as a means of blocking tumour growth
in many cancer models [6, 7]. Small molecule inhibitors of TNKS1 and TNKS2 have been
shown to reduce Wnt signalling in intestinal cancer cell lines and have been suggested to pre-
vent tumour growth due to their ability to stabilise AXIN1 and AXIN2 levels and as a result, to
inhibit β-catenin mediated transcription [5, 8–16]. However, the progression of TNKSi into
clinical trials has been restricted due to significant issues of intestinal toxicity within in vivo
models, emphasising the important role of Wnt signalling in adult tissue homeostasis [17]
[18]. Furthermore, the TNKSi described above reduced colorectal cancer cell numbers in 2D
culture, but did so with relatively low effect sizes by comparison with their ability to reduce lev-
els of Wnt/TCF-dependent transcription in the same cell lines. Similar partial-efficacy cell line
responses were observed during the development of inhibitors of the CDK8 and CDK19
kinases, suggesting that growth in 2D cell culture may not be an optimal readout for com-
pounds that are anticipated to target cancer stem cells [7].
To date, most preclinical in vitro studies have relied on the use of conventional 2D cultures
of cell lines. Whilst cancer cell lines can be usefully used to study pathway deregulation, they
frequently fail to be predictive when used as readouts of anti-tumour efficacy [19]. It is well
recognised that immortalised 2D cell lines adapted to growth on plastic poorly represent the
intricate cellular cross-talk present in tumours. In particular, they are unable to recapitulate
the inter-cellular interactions that form cancer stem cell niches and, as a consequence, such
models fail to fully reflect in vivo results.
3D tumour organoids have been shown to provide a more complex insight into tumour
cell-cell interactions [20, 21] and have been recognised as models with the potential to bridge
the gap between in vitro and in vivo preclinical studies. Organoids that have never been
adapted for growth on plastic have been cultured from multiple tumour types, and have been
shown to better represent the genetic diversity of distinct tumour subtypes than 2D cell lines
[22]. Organoids derived from genetically-engineered mice in which the Wnt pathway had
been oncogenically-activated accurately predicted subsequent in vivo responses to inhibitors
of the CDK8 and CDK19 kinases, while 2D cell culture only showed partial-efficacy [7]. Orga-
noids have been further shown to predict patient responses and might in future be used in the
clinic in personalised medicine [23].
The biological and phenotypic complexity of 3D primary organoid cultures allow, in princi-
ple, the assessment of compound activity against a subset of inter-cellular signalling that is cen-
tral to tumour growth. However, most uses of organoid assays to date have relied on fixed
end-point metabolic assays (e.g. ATP-level quantification) that aggregates responses in every
cell within a population of organoids. To fully exploit the potential of organoid assays, analysis
PLOS ONE
3D imaging of organoids identifies responses to Tankyrase inhibitors
PLOS ONE | https://doi.org/10.1371/journal.pone.0235319 August 18, 2020 2 / 20
https://www.cancerresearchuk.org Innovate UK
grant 9776 (KE) https://www.gov.uk/government/
organisations/innovate-uk. The funders did not play
any role in study design, data collection and
analysis, the decision to publish or the preparation
of the manuscript.
Competing interests: L.S.P is a founder and major
shareholder of OcellO B.V, and C.E., H.-P.B and D.
W are employees of Merck Healthcare KGaA. M.T
and J.T are employees of Cellesce Ltd.T.C.D is a
director of Cellesce Ltd. L.M.B and D.A.B are
siblings. The funder provided support in the form
of salaries for authors L.S.P, C.E.,H.-P.B, D.W, M.T
and J.T but did not have any additional role in the
study design, data collection and analysis, decision
to publish, or preparation of the manuscript. The
specific roles of these authors are articulated in the
‘author contributions’ section. This does not alter
our adherence to PLOS ONE policies on sharing
data and materials.
Abbreviations: CRC, Colorectal cancer; NOD/SCID,
Non-obese Diabetic/Severe Combined
Immunodeficiency; ODX, Organoid derived
xenograft; PARP, poly (ADP-ribose) polymerases;
TNKS, Tankyrase.

platforms need to yield quantitative data that can clearly represent compound-induced effects
on signalling pathways whose outputs are complex, whilst simultaneously maintaining usabil-
ity in a high throughput format.
In this work, we generated CRC patient-derived organoids and studied their responses to
Tankyrase inhibitors (TNKSi). Using metabolic end point assays, tankyrase inhibition showed
partial efficacy, reflecting a limited reduction in the growth of TNKSi-sensitive organoids.
Closer examination of organoid responses suggested that TNKSi altered the ratio of stem-like
to differentiated cell populations. This subtle effect on the cellular composition of organoids
was best reflected by multiparametric imaging analysis of organoids that was nonetheless com-
patible with high throughput analysis. Our findings demonstrate the potential of a phenotypic
approach to assess drug-induced phenotypes in a preclinical setting, which may be applicable
to a wider range of therapeutics that target cancer stem cell biology and niches.
Materials and methods
Materials
Corning Growth factor-reduced Matrigel was purchased from VWR (#734–1101). Cell culture
media were purchased from Invitrogen Life Technologies, and cell culture plastics from Nunc
unless otherwise stated. All TNKSi compounds were supplied and synthesized by Merck
Healthcare KGaA (Darmstadt, Germany). All stock solutions of compounds were reconsti-
tuted in DMSO. Activity of each compound had been assessed prior to shipping using a cell-
based immunobead assay to determine TNKSi potency based on the ability of the inhibitors to
stabilise the AXIN2 [24]. The determined EC
50
values of 2 nM, 36 nM and 319 nM (Fig 2A)
were determined for C1, C2, and C3 respectively.
Human tissue
Surgically resected patient materials were obtained from University Hospital of Wales by the
Wales Cancer Bank with written informed ethical consent from male and female patients
(>16 years of age, above the UK age of valid consent) with known or suspected malignant dis-
ease and anonymised (WCB project reference #12/001). The Wales Cancer Bank has ethics
approval as a Research Tissue Bank from the Wales Research Ethics Committee 3 (reference
16/WA/0256), and is licensed by the Human Tissue Authority under the UK Human Tissue
Act (2004) to store human tissue, taken from the living, for research (licence 12107) [25].
These approvals cover the collection of samples (including written consent), processing and
storing samples across multiple collection and storage sites. All patient derived material was
handled in concordance with HTA regulations. Histological sections of patient tissue were
imaged at University Hospital Wales and remained anonymised.
Organoid culture
The isolation of tumour organoids from patient material was processed as previously described
by Sato et al., [20] with some refinements outlined below. Briefly, following surgical resection,
tumour samples were obtained and stored in Hibernate A medium (Life Technologies) at 4˚C.
Tissue pieces were dissected to remove connective tissue and chopped to pieces approximately
2 mm in diameter. Following dissection, carcinoma pieces were washed in PBS then digested
enzymatically with Collagenase and Dispase for 30 minutes at 37˚C. Once digested, tissue was
triturated in PBS at room temperature to release cell fragments from tissue. Cell fragments
within the supernatant were then centrifuged for 5 min at 100 rcf, at 4˚C. Cell pellets were
resuspended in Advanced DMEM/F12 supplemented with 1% GlutaMAX, 1% HEPES buffer
PLOS ONE
3D imaging of organoids identifies responses to Tankyrase inhibitors
PLOS ONE | https://doi.org/10.1371/journal.pone.0235319 August 18, 2020 3 / 20

solution and 1% Penicillin/Streptomycin, then filtered (70 μm). Collected tissue fragments
were counted and plated at a density of 1000 cell fragments per 50 μL of Growth Factor
Reduced Matrigel (BD Biosciences) in 24 well plates (Nunc). Following Matrigel polymerisa-
tion, cells were overlaid with 500 μL of either “7+” or “Full” media and replenished every 4
days. ‘7+’ Media conditions consisted of Advanced DMEM/F12 supplemented with penicillin/
streptomycin, 2mM GlutaMAX, 10 mM HEPES,1 X B27 supplement, 1 X N2 supplement (all
Life Technologies), 1 mM N-acetyl-L-cysteine (Sigma-Aldrich). The following additional
niche factors were added to generate “Full” media: 50 ng/ml mouse recombinant EGF (Life
Technologies), 100 ng/ml mouse recombinant Noggin (Peprotech), 10% RSpo-1 conditioned
medium [26], 40% Wnt-3A conditioned medium [20], 500 nM A83-01 (Tocris) and 10 μM
SB202190 (Sigma). 10μM Y-27632 (Tocris) was also added to media for the first three days of
culture. Optimal media was determined for each tumour organoid line on the basis of condi-
tions that favoured the most efficient growth of organoids. For subsequent passaging, mainte-
nance in culture and analyses, the most favourable media condition was used. All culture were
maintained in humidified incubators at 37˚C, 5% CO
2.
DNA extraction
DNA extraction from organoid cultures was carried out using a QIAamp DNA Mini Kit (Qia-
gen) following the manufacturer’s instructions. Purified DNA quality was initially assessed
using a Nanodrop1000 (Thermo Fisher). Patient blood samples were processed by the Wales
Cancer Bank (WCB), genomic DNA was extracted from 4 ml of whole blood using the Chema-
gicSTAR automated cell lysis and DNA extraction workstation (Hamilton Company) housed
within the All Wales Medical Genetics Service (AWMGS) laboratory. Both DNA samples were
then passed to the Wales Gene Park for quality control, library preparation and sequencing.
Whole exome library preparation
The following was carried out by the Wales Gene Park. Genomic DNA was initially quantified
using the Qubit1 (Life Technologies) and the samples were serially diluted to 5 ng/μL. 50ng
of gDNA was used as the input and the sequencing libraries were prepared using the Illu-
mina1 Nextera Rapid Capture Enrichment kit (Illumina Inc.). The steps included tagmenta-
tion of gDNA, clean-up of the tagmented DNA, amplification, clean-up of the amplified DNA,
hybridisation of probes, capture of the hybridized probes, second hybridization of probes, sec-
ond capture, clean-up of the captured library, amplification of enriched library, clean-up of
the enriched library and finally validation of the complete library. The manufacturer’s instruc-
tions were largely followed with extra quantitation steps prior to the hybridization of the
probes to ensure that approximately 500 ng of each sample was pooled. The libraries were vali-
dated using the Agilent 2100 Bioanalyser and a high-sensitivity kit (Agilent Technologies) to
ascertain the insert size, and the Qubit1 (Life Technologies) was used to perform the fluoro-
metric quantitation. Following validation, the libraries were normalized to 10 nM, pooled
together and clustered on the cBot™2 following the manufacturer’s recommendations. Pools
were then sequenced using a 75-base paired-end (2x75bp PE) dual index read format on the
Illumina1 HiSeq4000 according to the manufacturer’s instructions.
Whole exome sequencing and mutation calling
Reads were mapped to the human reference genome (GRCh37) with BWA version 0.7.10 [27],
and duplicate reads were removed with Samtools version 0.1.19 [28]. Base quality score recali-
bration was performed using GATK’s BaseRecalibrator (GATK version 4.0.3.0, [29, 30] and
GATK’s Mutect2 was used to call somatic variants (tools found at https://software.
PLOS ONE
3D imaging of organoids identifies responses to Tankyrase inhibitors
PLOS ONE | https://doi.org/10.1371/journal.pone.0235319 August 18, 2020 4 / 20

broadinstitute.org/gatk/). Variants were annotated with Variant Effect Predictor version 91.3
[31], using a cached version of ensemble v91. A custom Perl script was used to extract genes of
interest from annotated variant files. Default parameters were used for all software unless oth-
erwise stated. All of the calls reported were cross-checked against the cBioportal database
(https://www.cbioportal.org) and are those found to have been previously described within the
COSMIC database (https://cancer.sanger.ac.uk/cosmic) as being “likely oncogenic”.
Immunohistochemistry
Formalin-fixed paraffin embedded histological sections of patient tissue were processed,
imaged and analysed by the Histopathology unit at University Hospital Wales. For organoid
histology, organoids embedded within Matrigel were fixed with 4% paraformaldehyde for 15
min at room temperature. Organoids were then lifted from wells and washed several times
with PBS prior to embedding within low melting-point agarose, followed by dehydration, par-
affin embedding, sectioning and Hematoxylin+Eosin (H&E) staining. Images were acquired
on an Olympus Dp26 microscope.
Western blot detection
Organoids were harvested from 6 X 50 μL blobs of Growth factor-reduced Matrigel using Cell
Recovery Solution (Corning) for 1 hour on ice. Total protein was extracted from organoids by
addition of 500 μL of lysis buffer to whole organoids (0.02 M Tris-HCl, 2 mM EDTA, 0.5% v/v
NP-40 (IPEGAL) in ddH2O containing 1x PhosSTOP phosphatase inhibitor (Roche) and 1x
complete Protease Inhibitor Cocktail (Roche)). The lysates were centrifuged at 8000 rcf for 15
min and proteins harvested in the supernatants. Protein content was measured using a BCA
protein assay kit (Pierce). Samples were resolved on Novex NuPAGE 4–12% Bis-Tris PAGE
gels, and then blotted onto nitrocellulose membranes using the Invitrogen iBlot Dry Blotting
cassette system.
Blots were blocked at room temperature for 1 h in 5% w/v milk powder in wash buffer, Tris
buffered saline (TBS) plus 0.2% Tween 20. Primary antibodies were diluted in blocking buffer
and incubated overnight at 4˚C. After washing the blots were incubated with HRP-conjugated
secondary antibodies for 1 h at room temperature. Bands were visualized using the enhanced
chemoluminescence (SuperSignal West Dura; Pierce). Antibodies and dilutions: TNKS1/2
(Santa Cruz (Tnks1/2 E-10) SC-365897, 1:500), Axin2 (Cell Signalling Technologies 76G6,
1:500) (# 2151)), β-actin (Sigma (AC-74) A2228, 1:10,000), anti-mouse or anti-rabbit HRP-
conjugated antibodies (GE Healthcare, 1:5000).
Organoid viability measurements
Organoids in culture were gently dissociated to near-single cell populations using TrypLE
(Life Technologies) before resuspension within growth factor-reduced Matrigel, and dispensed
into white clear bottomed 96 well plates in 9 μL Matrigel per well (400 cells/ μL of Matrigel).
Following Matrigel polymerization, a 9-point 2-fold dilution range of compounds were diluted
in tailored growth medium and dispensed within each well. DMSO was used as a negative con-
trol at a final concentration of 0.1%. Following a six day exposure to compounds, Cell Titer
Glo 3D (Promega) reagent was applied as per the manufacturer’s guidelines for an endpoint
readout. Relative Luminescence values were obtained on a BMG Fluostar plate reader. Dose
response curves and EC
50
values were obtained using Microsoft Excel with XLFit plug-in (Ver-
sion 5.4.0.8).
PLOS ONE
3D imaging of organoids identifies responses to Tankyrase inhibitors
PLOS ONE | https://doi.org/10.1371/journal.pone.0235319 August 18, 2020 5 / 20

Frequently asked questions

Q1. What are the contributions in "3d imaging of colorectal cancer organoids identifies responses to tankyrase inhibitors" ?

By contrast, 3D tumour organoids provide a platform in which complex cell-cell interactions can be studied. Using compounds with 3 orders of magnitude difference in cellular mechanistic potency together with image-based assays, the authors demonstrate that morphometric analyses can capture subtle alterations in organoid responses to Wnt inhibitors that are consistent with activity against a cancer stem cell subpopulation. Here, the authors generated tumour organoids from colorectal cancer patients and tested their responses to inhibitors of Tankyrase ( TNKSi ) which are known to modulate Wnt signalling. 

Q2. What was the method used to validate the libraries?

The libraries were validated using the Agilent 2100 Bioanalyser and a high-sensitivity kit (Agilent Technologies) to ascertain the insert size, and the Qubit1 (Life Technologies) was used to perform the fluorometric quantitation. 

Q3. What is the significance of the morphometric analysis of stem cell markers?

Alterations in the expression of stem cell marker genes including ASCL2 and LGR5 correlated with cellular responses, but larger-scale studies will be required to determine whether panels of gene expression markers reliably predict organoid (and ultimately patient) responses. 

Q4. What is the common use of organoids to date?

most uses of organoid assays to date have relied on fixed end-point metabolic assays (e.g. ATP-level quantification) that aggregates responses in every cell within a population of organoids. 

Q5. How many morphometric features were quantified using automated algorithms?

Numerous morphometric features, such as the distance of the matrix to central lumens and the number of nuclei per structure were quantified using automated algorithms. 

Q6. What is the way to assess the effects of TNKSi in a patient organ?

To assess whether both functional and phenotypic effects of TNKSi within the sensitive organoids would translate to a reduction in tumour growth in vivo, an organoid-derived xenograft model was generated from the TNKSi-sensitive line, Iso 75. 

Q7. What is the biomarker of organoid response?

The best overall biomarker of organoid response may instead be functional morphometric readouts, perhaps as a consequence of differential cellular responses to the compounds being studied. 

Q8. What were the steps for the enriched library?

The steps included tagmentation of gDNA, clean-up of the tagmented DNA, amplification, clean-up of the amplified DNA, hybridisation of probes, capture of the hybridized probes, second hybridization of probes, second capture, clean-up of the captured library, amplification of enriched library, clean-up of the enriched library and finally validation of the complete library. 

Q9. How did the TNKSi inhibitors affect the stemlike signature of organoid?

August 18, 2020 15 / 20suggest that modulation of Wnt signalling in a sensitive line was sufficient to limit the stemlike signature of organoid cell populations. 

Q10. What are the factors that may influence the morphological structure of the tumour?

Confounding factors in these analyses are likely to be the levels of stem cell gene expression, variation in the ‘stemness’ of tumours and the proportion of cancer stem cells within tumour and corresponding organoid populations. 

Q11. How many organoids were injected into the flanks of a healthy mouse?

Following treatment, small organoids were injected orthotopically into the flanks of 16gamma-irradiated Non-Obese Diabetic/Severe Combined Immunodeficiency (NOD/SCID/γ) mice, at one injection site per mouse (Fig 5A). 

Q12. What was the procedure for the validation of the libraries?

Following validation, the libraries were normalized to 10 nM, pooled together and clustered on the cBot™2 following the manufacturer’s recommendations. 

Q13. What is the way to predict phenotypic responses?

Taken together, these results suggest that the pharmacological inhibition of TNKS results in an overall reduction of Wnt/β-catenin signalling in both TNKSi-sensitive and resistant organoid lines and that phenotypic responses could not be simply predicted based on the expression of biomarkers. 

Q14. What was the effect size of the TNKSi-induced alterations?

Single parameter functional readouts such as TNKSi-induced alterations in ATP levels showed a reduced effect size by comparison with the morphometric responses. 

Q15. What is the morphometric analysis of Iso 38, Iso 72,?

The morphometric analysis showed that TNKS inhibition resulted in phenotypic alterationsin Iso 38, Iso 72, Iso 75 and Iso 78 (Fig 3B). 

Q16. What is the link between the loss of -catenin and TNKSi?

Taken together, this suggested that the loss of β-catenin was linked to changes in gene expression, but not necessarily to functional responses. 

Q17. What was the method used for the blotting of organoids?

Samples were resolved on Novex NuPAGE 4–12% Bis-Tris PAGE gels, and then blotted onto nitrocellulose membranes using the Invitrogen iBlot Dry Blotting cassette system.