Registered report: Wnt activity defines colon cancer stem cells and is regulated by the microenvironment

The Reproducibility Project: Cancer Biology seeks to address growing concerns about reproducibility in scientific research by replicating selected results from a substantial number of high-profile papers in the field of cancer biology. The papers, which were published between 2010 and 2012, were selected on the basis of citations and Altmetric scores (Errington et al., 2014). This Registered report describes the proposed replication plan of key experiments from ‘Wnt activity defines colon cancer stem cells and is regulated by the microenvironment’ by Vermeulen and colleagues, published in Nature Cell Biology in 2010 (Vermeulen et al., 2010). The key experiments that will be replicated are those reported in Figures 2F, 6D, and 7E. In these experiments, Vermeulen and colleagues utilize a reporter for Wnt activity and show that colon cancer cells with high levels of Wnt activity also express cancer stem cell markers (Figure 2F; Vermeulen et al., 2010). Additionally, treatment either with conditioned medium derived from myofibroblasts or with hepatocyte growth factor restored clonogenic potential in low Wnt activity colon cancer cells in vitro (Figure 6D; Vermeulen et al., 2010) and in vivo (Figure 7E; Vermeulen et al., 2010). The Reproducibility Project: Cancer Biology is a collaboration between the Center for Open Science and Science Exchange and the results of the replications will be published in eLife. DOI: http://dx.doi.org/10.7554/eLife.07301.001


Introduction
Wnt mediated activation of Frizzled receptors results in translocation of β-catenin to the nucleus where it binds to TCF/LEF transcription factors to induce expression of Wnt target genes. Wnt signaling proteins mediate a wide variety of biological processes during development, including maintenance of stem cell populations (Malanchi and Huelsken, 2009;Clevers et al., 2014), and aberrant Wnt activation is linked to several diseases, including cancer (Clevers and Nusse, 2012). Vermeulen and colleagues showed that high Wnt activity correlated with markers of colon cancer stem cells and enhanced clonogenic potential of cells (Vermeulen et al., 2010). They also showed that stromal myofibroblasts secreted factors such as HGF that enhanced Wnt activity and clonogenicity (Vermeulen et al., 2010). Further, treatment of more differentiated cells with myofibroblast conditioned medium (MFCM) enhanced Wnt activity in these cells and enhanced clonogenicity in vitro and in vivo, illustrating that colon cancer cell stemness can be modified by the microenvironment.
To assess Wnt activity in colon cancer cells, Vermeulen and colleagues utilized the TOP-GFP reporter system, a LEF-1/TCF responsive promoter driving expression of the enhanced GFP reporter (Reya et al., 2003). In Figure 2F, colon cancer stem cells were isolated from human colon cancer specimens and transduced with the TOP-GFP reporter. Wnt activity was then assessed using *For correspondence: tim@cos.io fluorescence-activated cell sorting (FACS) in populations derived from a single cell, and flow cytometry was concurrently used to assess the levels of the cancer stem cell markers CD133, CD24/ CD29 or CD44/CD166 in these cells. Vermeulen and colleagues found that cells with high Wnt activity levels correlated with expression of cancer stem cell markers (Vermeulen et al., 2010). Several other reports have used the TOP-GFP system or a similar reporter system to demonstrate that Wnt activity was enhanced in a population of cells expressing cancer stem cell markers. Correlation of cancer stem cells markers with high Wnt activity was found in primary and metastatic mouse mammary tumors (Malanchi et al., 2012), spheroid cultures of colon cancer cells (Colak et al., 2014), human and mouse colonic adenomas (Prasetyanti et al., 2013), and primary colon cancer cells (Kemper et al., 2012). All of these studies also found that cells with high Wnt activity maintained clonogenicity, while cells with low Wnt activity had markedly reduced clonogenic potential. In contrast, one report did find that while high Wnt reporter activity did correlate with expression of cancer stem cell markers, the tumorigenic capacity of cells was not dependent upon Wnt activity in four of the five cell lines tested, including three cell lines derived from primary colon cancers (Horst et al., 2012). The experiment presented in Figure 2F will be replicated in Protocol 2. Figure 6D assessed the potential of MFCM or HGF to enhance the clonogenic potential in vitro of colon cancer cells with low Wnt activity (TOP-GFP low cells). Clonogenic potential was measured using a limiting dilution assay in which cells were plated at a range of densities and the number of colonies that grow over time was counted. This experiment also examined the ability of the small molecular c-Met inhibitor, PHA-665752, to block MFCM or HGF-mediated clonogenicity. Vermeulen and colleagues showed that treatment of TOP-GFP low cells with HGF or MFCM increased clonogenicity of these poorly clonogenic cells, and that clonogenicity could be reversed by c-Met inhibition (Vermeulen et al., 2010). Additionally, MET has been reported to be enriched in glioblastoma stem cells (GSCs) and promote their self-renewal (Li et al., 2011;De Bacco et al., 2012;Joo et al., 2012). Met inhibition could also reduce the clonogenic and tumorigenic potential of GSCs, with activation of the Wnt/β-catenin signaling pathway shown to be a key mediator of the HGF/Met signaling pathway in these cells (Joo et al., 2012;Kim et al., 2013). This experiment is replicated in Protocol 3. This result was expanded upon in Figure 7E to assess the effect of MFCM on TOP-GFP low cells in vivo, where Vermeulen and colleagues found that TOP-GFP Low cells coinjected with MFCM, or an admixture of the factors these cells secrete, had enhanced tumorigenicity compared to TOP-GFP Low cells injected alone (Vermeulen et al., 2010). To assess the affect of MFCM on tumorigenicity, limiting dilutions of the different cell populations, in the presence or absence of MFCM, were injected subcutaneously into nude mice and tumor formation was measured over time. Likewise, HGF, along with other cytokines secreted from tumorassociated cells, was demonstrated to increase the tumorigenic activity and metastatic potential of colorectal cancer progenitor cells (Todaro et al., 2014). This experiment is replicated in Protocol 4.

Protocol 1: Isolation of colon cancer stem cells and infection with TOP-GFP
This experiment describes the isolation and culture of colon cancer stem cells. These spheroid cultures and the Co100 cell line used in the original study will be transduced with TOP-GFP and control plasmid. This will produce the single-cell-derived TOP-GFP clones used for further analysis in Protocols 2 and 3.

Sampling
■ This experiment will be performed once to generate two newly derived single spheroidal cultures, which along with the Co100 cell line will provide three colon cancer lines. ■ Each of the three lines will be used to generate and isolate 3 different TOP-GFP cancer stem cell (CSC) clones and one control clone. 1. Obtain twelve freshly excised human colon adenocarcinoma tissue fragments. a. If not enough viable spheroidal cultures are generated with the initial twelve fragments, an additional source of human colon adenocarcinoma cells will be obtained. b. Include histological diagnosis report and patient annotation. c. Note: with human colon tissue fragment samples, there is a 10-20% success rate of obtaining spheroidal cultures. 2. Wash 4 times in PBS supplemented with penicillin (500 U/ml), streptomycin (500 U/ml), and amphotericin B (1.25 μg/ml). 3. Incubate overnight in DMEM/F12 medium supplemented with penicillin (500 U/ml), streptomycin (500 U/ml), and amphotericin B (1.25 μg/ml). 4. Digest with collagenase (1.5 mg/ml) and hyaluronidase (20 μg/ml) in PBS for 1 hr at 37˚C; shake repeatedly during digestion. 5. Pass the dissociated sample through a 40 μm cell strainer and wash with CSC medium. 6. Remove erythrocytes and cell debris by Lympholyte-M centrifugation following manufacturer's instructions. 7. Wash cells 2-3 times with CSC medium and maintain culture. Maintain Co100 cell line following same methodology. a. Once a viable culture is established, colon cancer cells will cluster into spheroids (∼50-100 cells/ spheroid). b. Dissociate spheroids.

Materials and reagents
i. Pellet spheroids by centrifuging 5 min at 1000 RPM.
ii. Aspirate medium, being careful not to disrupt cell pellet.
iii. Using a sterile 5 ml serological pipet, gently resuspend cells in 3 ml of 1 mg/ml trypsin by pipetting up and down 3×. iv. Place tubes in 37˚C tissue culture incubator for 2.5 min. v. Agitate the cells by gently pipetting up and down 3 times using a sterile 5 ml serological pipet. vi. Return tubes to 37˚C tissue culture incubator for an additional 2.5 min. vii. Stop the dissociation by adding 10 ml of DMEM supplemented with 10% FBS. viii. Pellet cells by centrifuging 5 min at 1000 RPM. ix. Aspirate medium, being careful not to disrupt cell pellet.
x. Gently resuspend cells in an appropriate volume of CSC medium and perform a viable cell count. c. Cultures should be passaged when cell concentration exceeds 1 × 10 6 cells/ml of medium.
i. Additionaly, maintain cultures, untransduced, as a control for Protocols 2, 3, and 4. 8. Transduce the two newly derived dissociated single spheroidal cultures and the Co100 cell line lentivirally with the TOP-GFP construct following the Trono lab Protocol for 'Production of Lentiviral Vectors in 293T cells' briefly described with the following modifications. a. Plate 4.7-5.8 × 10 6 HEK293T cells per 15 cm plate. b. Transfect 15 cm plate with TOP-GFP lentiviral vector, packaging plasmid, and envelope plasmid with TransIT-293 transfection reagent following Trono lab protocol instructions. c. Change medium 6-8 hr later and add 15 ml/plate of fresh medium. d. Harvest 30 ml of medium per plate of transfected HEK293T cells (15 ml on day 1 and day 2 after transfection) and filter through a 0.45 μm filter. i. Store day 1 medium at 4˚C until it can be combined with day 2 harvest. e. Concentrate by centrifugation (O/N at 4˚C and 4000 RPM) in a 50 ml polypropylene conical tube. f. Resuspend virus pellet in 500 μl of Opti-MEM. g. Transduce dissociated spheroidal cultures with 20 μl of concentrated virus/1 × 10 6 cells in 10 ml of medium supplemented with 8 μg/ml polybrene. h. Change medium after 24 hr of infection to remove polybrene, dead cells, etc. 9. After 3-4 passages and 2 weeks in culture, or when cells are growing robustly, sort for single, propidium iodide-negative, and GFP-positive cells by FACS. a. Dissociate spheroids as described in step 7b. b. Add 250 ng/ml propidium iodide solution directly to cells before analysis. c. Single cells were gated within the GFP positive population. i. Note: No specific level of GFP was used in the original study. 10. Deposit 1 cell/well in a 96-well ultralow-adhesion plate containing CSC medium.

Reagent
a. Plate at least three plates with single cells as the incidence of visible spheres is around 1%. 11. After visible spheres arise, transfer to ultralow-adhesion plates and expand three independent TOP-GFP clones from the two newly derived spheroidal cultures and the Co100 cell line. a. Once spheroid cultures arise from a single cell clone, transfer the spheroid into one well of a 48 well plate and gently break down the spheroid by mechanical dissociation. b. Slowly scale up to a larger culture surface (i.e., 1 well of a 48 well plate, then 4 wells of a 24 well plate, then 1 well of a 6 well plate, then 25 cm 2 flask, etc) by dissociating the culture (as described in step 7b) for the subsequent experiments (Protocols 2, 3, and 4). 12. Maintain clones from single-cell-derived TOP-GFP cells and untransduced parental spheroidal cultures, for further analysis (Protocols 2, 3, and 4).

Deliverables
■ Data to be collected: ○ Histological diagnosis and patient annotation of human colon tissue fragments. ○ All FACS plots in gating scheme (including controls), leading to final population of single, propidium iodide-negative, GFP-positive cells. ■ Sample delivered for further analysis: ○ Spheroidal cultures (two newly derived and the Co100 cell line) as a control in Protocols 2, 3, and 4. ○ Single-cell-derived TOP-GFP clones (two newly derived and the Co100 cell line) for further analysis in Protocols 2, 3, and 4.

Confirmatory analysis plan
N/A.

Known differences from the original study
All known differences are listed in the materials and reagents section above with the originally used item listed in the comments section. All differences have the same capabilities as the original and are not expected to alter the experimental design.

Provisions for quality control
A report of histological diagnosis and patient annotation will be included. Additionally, cell viability will be monitored during culture conditions. All of the raw data, including the FACS plots, will be uploaded to the project page on the OSF (https://osf.io/pgjhx) and made publically available.
Protocol 2: flow cytometry analysis of CSC marker expression in TOP-GFP clones This experiment will assess the association of TOP-GFP levels with CSC marker expression, specifically CD133, CD29, CD24, CD44, and CD166, which is a replication of the experiment reported in Figure 2F.

Sampling
■ This experiment will be performed with each of the 3 different TOP-GFP CSC clones from the two newly derived cultures and the Co100 cell line. ■ Each TOP-GFP CSC clone will be analyzed for signal intensity for a total power of ≥ 80%. Deliverables ■ Data to be collected: ○ All FACS plots in gating scheme (including controls), leading to final population of propidium iodide-negative, GFP-positive cells for analysis of each CD marker. ○ FACS mean fluorescence intensity and confidence intervals for each CD marker.

Confirmatory analysis plan
This replication attempt will perform the following statistical analysis listed below and compute the effects sizes for each TOP-GFP CSC clone.
■ Statistical Analysis: Note: 1. Since these tests will be performed for each of the three clones from the CSC cultures the alpha error will be adjusted with the Bonferroni correction. ○ Unpaired, two-tailed t-test with the Bonferroni correction for multiple comparisons: ■ CD133 expression from TOP-GFP low cells compared to TOP-GFP hgh cells.
■ CD24 expression from TOP-GFP low cells compared to TOP-GFP hgh cells. ■ CD29 expression from TOP-GFP low cells compared to TOP-GFP hgh cells. ■ CD44 expression from TOP-GFP low cells compared to TOP-GFP hgh cells. ■ CD166 expression from TOP-GFP low cells compared to TOP-GFP hgh cells.
■ Meta-analysis of effect sizes: ○ Compare the effect sizes of the TOP-GFP CSC clones from the two newly derived cultures and the Co100 cell line (3 different clones each) and use a meta-analytic approach to combine the effects, which will be presented as a forest plot.
Known differences from the original study All known differences are listed in the materials and reagents section above with the originally used item listed in the comments section. All differences have the same capabilities as the original and are not expected to alter the experimental design.

Provisions for quality control
Negative staining, individual antibody, and isotype controls are included to assess antibody staining relative to background. All of the raw data, including the FACS plots, will be uploaded to the project page on the OSF (https://osf.io/pgjhx) and made publically available.

Protocol 3: clonogenicity assay of TOP-GFP CSC clones
This experiment will assess the effect of MFCM and recombinant HGF on the clonogenic potential of the TOP-GFP CSC clones. This experiment will also examine the ability of the small molecular c-Met inhibitor, PHA-665752, to block MFCM-or HGF-triggered clonogenicity. This is a replication of the experiment reported in Figure 6D.

Sampling
■ This experiment will be performed with each of the 3 different TOP-GFP CSC clones from the two newly derived cultures and the Co100 cell line and with each cohort assessing 96 wells for a total power of ≥ 80%. ○ See 'Power calculations' section for details. ■ Each experiment has 8 cohorts: ○ Cohort 1: TOP-GFP low cells. ■ The titration of cells might need to be adjusted depending on the clonogenic potential of the spheroid clones. An initial pilot experiment will be performed to assess the potential for the three populations (TOP-GFP low , TOP-GFP high , and total TOP-GFP), without treatment, before proceeding with this design.

Note:
c TOP-GFP CSC clones, and untransduced spheroidal culture (no GFP) control, are generated in Protocol 1. c CSC cultures maintained in CSC medium: modified neurobasal A medium supplemented with 1× N2 supplement, lipid mixture-1 (1 ml/500 ml medium), basic fibroblast growth factor (20 ng/ml), and epidermal growth factor (50 ng/ml) at 37˚C in a humidified atmosphere at 5% CO 2 . c 18Co cells maintained in DMEM supplemented with 10% FBS and 1% glutamine at 37˚C in a humidified atmosphere at 5% CO 2 . c 18Co cells will be sent for mycoplasma testing and STR profiling. c An initial pilot experiment, performed once for each clone, will be performed to assess the clonogenic potential of the spheroid cultures. This will be performed with TOP-GFP low , TOP-GFP high , and total TOP-GFP gated populations left untreated (that is they will not be treated after being deposited in step 3 below). Depending on the outcome, the titration of cells might need to be adjusted before proceeding with the entire experiment as described.
1. After obtaining single cell suspensions in Protocol 1, dissociate TOP-GFP CSC clones and untransduced cultures with trypsin as described in Protocol 1 and resuspend 2 × 10 6 cells/ml in 500 μl of CSC medium for sorting. 2. Sort for single, propidium iodide-negative, and GFP-positive cells by FACS.
a. Add 250 ng/ml propidium iodide solution directly to cells before analysis.

Confirmatory analysis plan
This replication attempt will perform the following statistical analysis listed below and compute the effects sizes for each TOP-GFP CSC clone.
■ Statistical Analysis: Note: 1. Extreme limiting dilution analysis (ELDA) will be used to perform these tests (Hu and Smyth, 2009). 2. Since these tests will be performed for each of the three clones from the CSC cultures the alpha error will be adjusted with the Bonferroni correction. ○ Chi-square test for differences between any of the groups.
■ Planned pairwise differences between groups with the Bonferroni correction for multiple comparisons: 1. TOP-GFP low cells compared to TOP-GFP hgh cells. 2. TOP-GFP low cells compared to TOP-GFP low cells with HGF. 3. TOP-GFP low cells compared to TOP-GFP low cells with MFCM. 4. TOP-GFP low cells with HGF compared to TOP-GFP low cells with HGF and PHA-665752. 5. TOP-GFP low cells with MFCM compared to TOP-GFP low cells with MFCM and PHA-665752. 6. TOP-GFP whole cells compared to TOP-GFP whole cells with PHA-665752. ■ Meta-analysis of effect sizes: ○ Compute the effect sizes of the TOP-GFP CSC clones from the two newly derived cultures and the Co100 cell line (3 different clones each), compare them against the effect size in the original paper, and use a meta-analytic approach to combine the original and replication effects, which will be presented as a forest plot.
Known differences from the original study All known differences are listed in the materials and reagents section above with the originally used item listed in the comments section. All differences have the same capabilities as the original and are not expected to alter the experimental design.

Provisions for quality control
The cell line used in this experiment will undergo STR profiling to confirm its identity and will be sent for mycoplasma testing to ensure there is no contamination. Negative staining and untransduced spheroidal culture are included for gating. An initial experiment, performed once for each clone, will be performed to assess the clonogenic potential of the spheroid cultures to ensure the titration curves are appropriate. The amount of HGF in MFCM will be determined by ELISA to determine 18Co cells are producing HGF. All of the raw data, including the FACS plots, will be uploaded to the project page on the OSF (https://osf.io/pgjhx) and made publically available.
Protocol 4: effect of MFCM on tumorigenicity in TOP-GFP low CSC clone This experiment will assess the effect of MFCM on the tumorigenicity potential of one of the TOP-GFP CSC clones, which is a replication of Figure 7E.

Sampling
■ This experiment will be performed with one of the TOP-GFP CSC clones. ○ The clone used in this experiment will be from one of the two newly derived cultures with the largest difference in TOP-GFP low and TOP-GFP high as determined from Protocol 3 with untreated cells. If none of the three clones from either culture have differences similar to the reported values in the original study, then the clone with the largest difference from the Co100 cell line will be used. ■ Experiment will be performed with 4 mice per injection (a total of 16 mice per cohort) for a total power of ≥ 88%. ○ See 'Power calculations' section for details. ■ Each experiment has 3 cohorts: ○ Cohort 1: TOP-GFP low cells injected into nude mice. c 10, 100, 1000, and 5000 cells injected. ○ Cohort 2: TOP-GFP high cells injected into nude mice. c 10, 100, 1000, and 5000 cells injected. ○ Cohort 3: TOP-GFP low cells + MFCM injected into nude mice. c 10, 100, 1000, and 5000 cells injected. Deliverables ■ Data to be collected:

Materials and reagents
○ STR profile and result of mycoplasma testing of 18Co cells. ○ Raw data, standard curve, and concentration of HGF in MFCM. ○ All FACS plots in gating scheme (including controls), leading to final population of propidium iodide-negative, GFP-positive cells. ○ Mouse health records (including when tumors become detectable and caliper measurement).

Confirmatory analysis plan
This replication attempt will perform the following statistical analysis listed below.
■ Statistical Analysis: Note: 1. ELDA will be used to perform these tests (Hu and Smyth, 2009).
○ Chi-square test for differences between any of the groups. ■ Planned pairwise differences between groups with the Bonferroni correction for multiple comparisons: 1. TOP-GFP low cells compared to TOP-GFP hgh cells.

TOP-GFP low cells compared to TOP-GFP low cells with MFCM.
■ Meta-analysis of effect sizes: ○ Compute the effect sizes of each comparison, compare them against the effect size in the original paper and use a meta-analytic approach to combine the original and replication effects, which will be presented as a forest plot.

Known differences from the original study
The experiment requires transporting the cells from a facility that is sorting the cells to another facility (∼30 min away) to inject and monitor the mice. The replication will not include the TOP-GFP intermediate or TOP-GFP low with myofibroblasts that were reported for the C100.G7 clone. All known differences are listed in the materials and reagents section above with the originally used item listed in the comments section. All differences have the same capabilities as the original and are not expected to alter the experimental design.

Provisions for quality control
The cell line used in this experiment will undergo STR profiling to confirm its identity and will be sent for mycoplasma testing to ensure there is no contamination. Additionally, cells will be tested against a rodent pathogen panel to ensure no contamination by pathogens prior to implantation into nude mice. The amount of HGF in MFCM will be determined by ELISA to determine 18Co cells are producing HGF. Negative staining and untransduced spheroidal culture are included for gating. Confirmation of tumor incidence will be confirmed at the end of the study by necropsy. All of the raw data, including the FACS plots, will be uploaded to the project page on the OSF (https://osf.io/pgjhx) and made publically available.

Power calculations
For additional details on power calculations, please see analysis scripts and associated files on the Open Science Framework: https://osf.io/rfuj2/.

Protocol 1
Not applicable.

Protocol 2
Original data: unavailable and unable to be estimated.

Test family
■ 2 tailed t test, difference between two independent means: Bonferroni correction: alpha error = 0.003333 (corrected for the three clones from the CSC cultures and the multiple comparisons listed below).

Protocol 3
Summary of original data (estimated and simulated from Figure 6D) performed with R software, version 3.1.2 (R Development Core Team, 2014).
c The estimated stem cell frequency and 95% lower confidence interval were used to create simulated data sets with preserved sampling structure using ELDA (Hu and Smyth, 2009

Protocol 4
Summary of original data (obtained from Figure 7E) performed with R software, version 3.1.2 (R Development Core Team, 2014).
c The estimated stem cell frequency and 95% lower confidence interval were used to create simulated data sets with preserved sampling structure using ELDA (Hu and Smyth, 2009). c Both clones reported in Figure 7E were used to determine sample size to ensure an adequate number of mice are used to detect either effect size.

Test family
■ Chi-square test, differences between any of the groups: alpha error = 0.05.
Power calculations performed with R software, version 3. Summary of original data (obtained from Figure 7E) performed with R software, version 3.1.2 (R Development Core Team, 2014).

Test family
■ Chi-square test, differences between any of the groups: alpha error = 0.05.
Power calculations performed with R software, version 3.1.2 (R Development Core Team, 2014).