1. Developmental Biology and Stem Cells
Download icon

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

  1. James Evans
  2. Anthony Essex
  3. Hong Xin
  4. Nurith Amitai
  5. Lindsey Brinton
  6. Erin Griner
  7. Reproducibility Project: Cancer Biology Is a corresponding author
  1. PhenoVista Biosciences, California
  2. Explora BioLabs, California
  3. University of Virginia, Virginia
Registered Report
Cited
4
Views
1,092
Comments
0
Cite as: eLife 2015;4:e07301 doi: 10.7554/eLife.07301

Abstract

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.

https://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 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-GFPlow 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-GFPlow 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-GFPlow cells in vivo, where Vermeulen and colleagues found that TOP-GFPLow cells coinjected with MFCM, or an admixture of the factors these cells secrete, had enhanced tumorigenicity compared to TOP-GFPLow 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 tumor-associated 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.

Materials and methods

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.

Materials and reagents

ReagentTypeManufacturerCatalog #Comments
Modified neurobasal A mediumCell cultureLife Technologies10,888-022Original catalog # not specified
N2 supplementCell cultureLife Technologies17,502-048Original catalog # not specified
Lipid mixture-1Cell cultureSigma–AldrichL0288Original catalog # not specified
Fibroblast growth factor—Basic, human (FGF)Growth factorSigma–AldrichF0291Original brand not specified
Epidermal growth factor, human (EGF)Growth factorSigma–AldrichE9644Original brand not specified
Human colon tissue fragmentsClinical sampleN/AN/AOriginal mainly used microsatellite stable primary tumors.
Phosphate buffered saline (PBS) without MgCl2 and CaCl2BufferSigma–AldrichD8537Original brand not specified*
100× Penicillin/streptomycinCell cultureSigma–AldrichP4333Original brand not specified*
Amphotericin BCell cultureSigma–AldrichA2942Replaces GIBCO brand used in original study*
DMEM/F12 medium with L-glutamine and sodium bicarbonateCell cultureSigma–AldrichD8062Replaces GIBCO brand used in original study*
Collagenase, type 1A-SCell cultureSigma–AldrichC9722Replaces Roche brand used in original study
Hyaluronidase from bovine testes, type 1-S, 400–1000 units/mgCell cultureSigma–AldrichH3506Original brand not specified
40 µm cell strainerLabwareCorning431750Original brand not specified
Lympholyte-MChemicalCedarlaneCL5031Original catalog # not specified
Co100 cultureCell cultureAuthorsN/AFrom original lab
HEK293T cellsCell lineATCCCRL-3216Included during communication with authors. Original brand not specified
Dulbecco's Modified Eagle's Medium (DMEM)—high glucoseCell cultureSigma–AldrichD6429Originally not specified
Fetal Bovine Serum (FBS)Cell cultureSigma–AldrichF0392Originally not specified
0.5% trypsin/0.48 mM EDTA (5 mg/ml)Cell cultureSigma–AldrichT4174Included during communication with authors. Original brand not specified
150 mm tissue culture platesLabwareCorning430599Included during communication with authors. Original brand not specified
100 mm tissue cultures platesLabwareCorning430167Originally not specified
TOP-CMV-GFP reporter lentivirus vectorDNA constructAuthorsN/AFrom original lab
psPAX2 packaging plasmidDNA constructAuthorsN/AFrom original lab
pMD2.G envelope plasmidDNA constructAuthorsN/AFrom original lab
TransIT-293Transfection reagentMirus BioMIR 2704Originally not specified
50 ml polypropylene conical tubesLabwareCorning430290Included during communication with authors. Original brand not specified
OptiMEM-1 reduced serum mediumCell cultureLife Technologies31985062Included during communication with authors.
Hexadimethrine bromide (polybrene)Cell cultureSigma–Aldrich107689Included during communication with authors. Original brand not specified
Propidium iodideChemicalSigma–AldrichP4170Original brand not specified
FACS sorterEquipmentBD BiosciencesFACSaria
96-well ultralow-attachment plate, flat bottomLabwareCorning3474Original catalog # not specified
24-well ultralow-attachment plateLabwareCorning3473Included during communication with authors.
6-well ultralow-attachment plateLabwareCorning3471Included during communication with authors.
25 cm2 ultralow-attachment tissue culture flaskLabwareCorning3815Original catalog # not specified
75 cm2 ultralow- attachment tissue culture flaskLabwareCorning3814Original catalog # not specified
  1. *

Procedure

Note:

  • This protocol contains information described in Todaro et al., 2007.

  • Fresh primary cells and Co100 cell line maintained in CSC medium: modified neurobasal A medium supplemented with 1× N2 supplement, lipid mixture-1 (1 ml/500 ml medium), fibroblast growth factor-basic (20 ng/ml), and epidermal growth factor (50 ng/ml) at 37°C in a humidified atmosphere at 5% CO2.

  • HEK293T cells maintained in DMEM supplemented with 10% FBS at 37°C with 5% CO2.

  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.

      • 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 × 106 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 × 106 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 × 106 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.

    • 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 cm2 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%.

    • ○ See ‘Power calculations’ section for details.

  • ■ Staining conditions for each clone:

    • ○ CD133.

    • ○ CD24 and CD29.

    • ○ CD24 alone.

    • ○ CD29 alone.

    • ○ CD44 and CD166.

    • ○ CD44 alone.

    • ○ CD166 alone.

    • ○ Isotype controls.

    • ○ Unstained control.

    • ○ Untransduced spheroidal culture (no GFP) control.

Materials and reagents

ReagentTypeManufacturerCatalog #Comments
0.05% trypsin/0.48 mM EDTACell cultureSigma–AldrichT3924Original brand not specified
Phosphate buffered saline (PBS) without MgCl2 and CaCl2BufferSigma–AldrichD8537Originally not specified
Bovine serum albumin (BSA)ChemicalSigma–AldrichA3803Included during communication with authors. Original brand not specified
CD133 (clone AC133) -PE antibody (mouse IgG1)AntibodiesMiltenyi Biotec130-098-826Use at 1:100 dilution. Conjugate selected by replicating lab.
CD44 (clone G44-26)—APC antibody (mouse IgG2b, κ)AntibodiesBD Biosciences560890Use at 1:100 dilution. Conjugate selected by replicating lab.
CD166 (clone 105902)—PE antibody (mouse IgG1)AntibodiesR&D SystemsFAB6561POriginal clone listed as 105901. Use at 1:100 dilution. Conjugate selected by replicating lab.
CD24 (clone ML5)—PE antibody (mouse IgG2a, κ)AntibodiesBD Biosciences560991Use at 1:100 dilution. Conjugate selected by replicating lab.
CD29 (clone MAR4)—APC antibody (mouse IgG1, κ)AntibodiesBD Biosciences561794Use at 1:100 dilution. Conjugate selected by replicating lab.
Anti-IgG1—PE, mouse (clone X-56)AntibodiesMilitenyi Biotec130-098-106Use at 1:100 dilution. Originally not specified.
Anti-IgG2b κ—APC, mouse (clone 27–35)AntibodiesBD Biosciences555745Use at 1:100 dilution. Originally not specified.
Anti-IgG2a κ—PE, mouse (clone G155-178)AntibodiesBD Biosciences555574Use at 1:100 dilution. Originally not specified.
Anti-IgG1 κ—APC, mouse (clone MOPC-21)AntibodiesBD Biosciences555751Use at 1:100 dilution. Originally not specified.
Propidium iodideChemicalSigma–AldrichP4170Original brand not specified
FACS sorterEquipmentBD BiosciencesFACSaria

Procedure

Note:

  • TOP-GFP CSC clones, and untransduced spheroidal culture (no GFP) control, are generated in Protocol 1.

  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 re-suspend 1 × 106 cells/ml in PBS supplemented with 1% BSA.

  2. Stain cells with the following antibodies:

    • a. CD133-PE (use at 1:100 dilution).

    • b. CD24-PE (use at 1:100 dilution) and CD29-APC (use at 1:100 dilution).

    • c. CD44-APC (use at 1:100 dilution) and CD166-PE (use at 1:100 dilution).

    • d. Incubate antibodies with cells for 10 min in a dark refrigerator at 4°C (2–8°C is acceptable).

    • e. Wash cells by adding 20× the reaction volume of PBS with 1% BSA and gently inverting tubes 3× (i.e., cell/antibody volume is 100 μl, add 2 ml PBS supplemented with 1% BSA). Centrifuge cells at 1000 RPM for 10 min. Carefully aspirate supernatant completely. Resuspend cells in 100 μl PBS.

    • f. Include an unstained control for gating.

    • g. Include untransduced spheroidal culture (no GFP) for gating.

    • h. Include isotype control antibody stains.

      • i. Anti-IgG1—PE.

      • ii. Anti-IgG2b κ—APC.

      • iii. Anti-IgG2a κ—PE.

      • iv. Anti-IgG1 κ—APC.

  3. Add 250 ng/ml propidium iodide solution to cells just before analysis.

    • a. Include an unstained control for gating.

  4. Perform flow cytometry analysis for the following populations:

    • a. Analyze CD133 intensity:

      • i. Gate for viable cells (propidium iodide-negative cells).

      • ii. Gate for TOP-GFP expression.

        1. Identify top 10% and bottom 10% of TOP-GFP expression.

        2. Analyze CD133 intensity in at least 10,000 cells in each fraction.

      • iii. Gate against a negative control (unstained cells).

    • b. Analyze CD24 and CD29 intensity:

      • i. Gate for viable cells (propidium iodide-negative cells).

      • ii. Gate for TOP-GFP expression.

        1. Identify top 10% and bottom 10% of TOP-GFP expression.

        2. Analyze CD29 and CD24 intensity in at least 10,000 cells in each fraction.

      • iii. Gate against a negative control (unstained cells) and cells stained with each antibody individually.

    • c. Analyze CD44 and CD166 intensity:

      • i. Gate for viable cells (propidium iodide-negative cells).

      • ii. Gate for TOP-GFP expression.

        1. Identify top 10% and bottom 10% of TOP-GFP expression.

        2. Analyze CD44 and CD166 intensity in at least 10,000 cells in each fraction.

      • iii. Gate against a negative control (unstained cells) and cells stained with each antibody individually.

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-GFPlow cells compared to TOP-GFPhgh cells.

      • ■ CD24 expression from TOP-GFPlow cells compared to TOP-GFPhgh cells.

      • ■ CD29 expression from TOP-GFPlow cells compared to TOP-GFPhgh cells.

      • ■ CD44 expression from TOP-GFPlow cells compared to TOP-GFPhgh cells.

      • ■ CD166 expression from TOP-GFPlow cells compared to TOP-GFPhgh 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-GFPlow cells.

    • ○ Cohort 2: TOP-GFPhigh cells.

    • ○ Cohort 3: TOP-GFPlow cells + HGF.

    • ○ Cohort 4: TOP-GFPlow cells + MFCM.

    • ○ Cohort 5: TOP-GFPlow cells + HGF + PHA-665752.

    • ○ Cohort 6: TOP-GFPlow cells + MFCM + PHA-665752.

    • ○ Cohort 7: total TOP-GFP cells.

    • ○ Cohort 8: total TOP-GFP cells + PHA-665752.

  • ■ Each cohort plates (per 96 well plate):

    • ○ 1 cell × 24 wells.

    • ○ 2 cells × 16 wells.

    • ○ 4 cells × 8 wells.

    • ○ 8 cells × 8 wells.

    • ○ 16 cells × 8 wells.

    • ○ 32 cells × 8 wells.

    • ○ 64 cells × 8 wells.

    • ○ 128 cells × 8 wells.

    • ○ 256 cells × 8 wells.

      • ■ 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-GFPlow, TOP-GFPhigh, and total TOP-GFP), without treatment, before proceeding with this design.

Materials and reagents

ReagentTypeManufacturerCatalog #Comments
Modified neurobasal A mediumCell cultureLife Technologies10,888-022Original catalog # not specified
N2 supplementCell cultureLife Technologies17,502-048Original catalog # not specified
Lipid mixture-1Cell cultureSigma–AldrichL0288Original catalog # not specified
Fibroblast growth factor—Basic humanCell cultureSigma–AldrichF0291Original brand not specified
Epidermal growth factor humanCell cultureSigma–AldrichE9644Original brand not specified
Phosphate buffered saline (PBS) without MgCl2 and CaCl2BufferSigma–AldrichD8537Original brand not specified
0.05% trypsin/0.48 mM EDTACell cultureSigma–AldrichT3924Originally not specified
18Co cellsCellsATCCCRL-1459
Dulbecco's Modified Eagle's Medium (DMEM)—high glucoseCell cultureSigma–AldrichD5671Original brand not specified
Fetal Bovine Serum (FBS)Cell cultureSigma–AldrichF0392Original brand not specified
L-glutamineCell cultureSigma–AldrichG7513Original brand not specified
Propidium iodideChemicalSigma–AldrichP4170Original brand not specified
FACS sorterEquipmentBD BiosciencesFACSaria
96-well ultralow-attachment plate, flat bottomLabwareCorning3474Original catalog # not specified
DMSOChemicalSigma–AldrichD8418Originally not specified
PHA-665752InhibitorSigma–AldrichPZ0147Replaces Pfizer brand used in original study
HGF, humanGrowth factorSigma–AldrichH5791Replaces Relia Tech. Inc. brand used in original study
T75 flaskLabwareCorning430641UOriginally not specified
Human HGF ELISA kitKitSigma–AldrichRAB0212Included for additional quality control measure
Plate reader capable of measuring absorbance at 450 nmInstrumentUsed for HGF ELISA
Hermes WiScan microscopeInstrumentIDEA Bio-MedicalOriginally not specified

Procedure

Note:

  • TOP-GFP CSC clones, and untransduced spheroidal culture (no GFP) control, are generated in Protocol 1.

  • 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% CO2.

  • 18Co cells maintained in DMEM supplemented with 10% FBS and 1% glutamine at 37°C in a humidified atmosphere at 5% CO2.

  • 18Co cells will be sent for mycoplasma testing and STR profiling.

  • 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-GFPlow, TOP-GFPhigh, 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 × 106 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.

      • i. Include an unstained control for gating.

      • ii. Include untransduced spheroidal culture (no GFP) for gating.

    • b. Gate for top 10% and bottom 10% for TOP-GFP expression.

    • c. With an additional sample gate for total TOP-GFP-positive cells.

  3. Deposit cells from TOP-GFPlow, TOP-GFPhigh, or total TOP-GFP cell populations into 96-well ultralow-adhesion plates with 100 µl of one type of medium added to each plate. Plate at 1 cell × 24 wells, 2 cells × 16 wells, 4 cells × 8 wells, 8 cells × 8 wells, 16 cells × 8 wells, 32 cells × 8 wells, 64 cells × 8 wells, 128 cells × 8 wells, and 256 cells × 8 wells per 96-well plate. The following medium conditions are used:

    • a. CSC medium.

    • b. CSC medium + 500 nM PHA-665752.

    • c. CSC medium + 25 ng/ml HGF.

    • d. CSC medium + MFCM.

    • e. CSC medium + 25 ng/ml HGF + 500 nM PHA-665752.

    • f. CSC medium + MFCM + 500 nM PHA-665752.

      • i. Prepare MCFM before treatment as follows:

        1. Seed 7.5 × 105 18Co cells in 75-cm2 flasks and incubate overnight.

        2. The next day wash cells twice with PBS and incubate for 24 hr with 10 ml of CSC medium without EGF and FGF-basic.

        3. The next day collect MFCM and clear by centrifugation for 5 min at 1400 RPM.

        4. Use at 1:2 dilution in CSC medium.

        5. The level of HGF present in MFCM will be determined by an ELISA following manufacturer's instructions.

  4. Incubate at 37°C for ∼10 days until evaluation of clonogenic potential.

    • a. Replace with the appropriate culture medium condition every 4 days during incubation.

  5. Count the number of cultures with spheres formed by bright field or GFP fluorescence microscopy.

    • a. Exclude any contaminated cultures from analysis.

Deliverables

  • ■ Data to be collected:

    • ○ 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.

    • ○ Raw counts of spheroidal cultures and total cultures examined (including if any wells were excluded).

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-GFPlow cells compared to TOP-GFPhgh cells.

    2. TOP-GFPlow cells compared to TOP-GFPlow cells with HGF.

    3. TOP-GFPlow cells compared to TOP-GFPlow cells with MFCM.

    4. TOP-GFPlow cells with HGF compared to TOP-GFPlow cells with HGF and PHA-665752.

    5. TOP-GFPlow cells with MFCM compared to TOP-GFPlow cells with MFCM and PHA-665752.

    6. TOP-GFPwhole cells compared to TOP-GFPwhole 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-GFPlow 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-GFPlow and TOP-GFPhigh 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-GFPlow cells injected into nude mice.

      • 10, 100, 1000, and 5000 cells injected.

    • ○ Cohort 2: TOP-GFPhigh cells injected into nude mice.

      • 10, 100, 1000, and 5000 cells injected.

    • ○ Cohort 3: TOP-GFPlow cells + MFCM injected into nude mice.

      • 10, 100, 1000, and 5000 cells injected.

Materials and reagents

ReagentTypeManufacturerCatalog #Comments
Modified neurobasal A mediumCell cultureLife technologies10,888-022Original catalog # not specified
N2 supplementCell cultureLife technologies17,502-048Original catalog # not specified
Lipid mixture-1Cell cultureSigma–AldrichL0288Original catalog # not specified
Fibroblast growth factor—Basic humanCell cultureSigma–AldrichF0291Original brand not specified
Epidermal growth factor humanCell cultureSigma–AldrichE9644Original brand not specified
Phosphate buffered saline (PBS) without MgCl2 and CaCl2BufferSigma–AldrichD8537Original brand not specified
0.05% trypsin/0.48 mM EDTACell cultureSigma–AldrichT3924Originally not specified
18Co cellsCellsATCCCRL-1459
Dulbecco's Modified Eagle's Medium (DMEM)—high glucoseCell cultureSigma–AldrichD5671Original brand not specified
Fetal Bovine Serum (FBS)Cell cultureSigma–AldrichF0392Original brand not specified
L-glutamineCell cultureSigma–AldrichG7513Original brand not specified
T75 flaskLabwareCorning430641UOriginally not specified
Human HGF ELISA kitKitSigma–AldrichRAB0212Included for additional quality control measure
Plate reader capable of measuring absorbance at 450 nmInstrumentUsed for HGF ELISA
Propidium iodideChemicalSigma–AldrichP4170Original brand not specified
FACS sorterEquipmentBD BiosciencesFACSaria
96-well ultralow-attachment plate, flat bottomLabwareCorning3474Original catalog # not specified
Growth factor reduced MatrigelCell cultureCorning356230Original brand not specified
8–15 week old female athymic nude mice (20–30 grams)Animal modelHarlanHsd:Athymic Nude-Foxn1nu
27 ½ G needleLabwareBD Biosciences305109Original not specified
ACE 1 ml Luer Lock syringeLabwareBD Biosciences309628Original not specified

Procedure

Note:

  • TOP-GFP CSC clone is generated in Protocol 1.

  • 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% CO2.

  • 18Co cells maintained in DMEM supplemented with 10% FCS and 1% glutamine at 37°C in a humidified atmosphere at 5% CO2.

  • 18Co cells will be sent for mycoplasma testing and STR profiling.

  1. After obtaining single cell suspensions in Protocol 1, dissociate TOP-GFP CSC clone and untransduced culture with trypsin as described in Protocol 1 and resuspend 2 × 106 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.

      • i. Include an unstained control for gating.

      • ii. Include untransduced spheroidal culture (no GFP) for gating.

    • b. Gate for top 10% and bottom 10% for TOP-GFP expression.

  3. Deposit cells from TOP-GFPlow and TOP-GFPhigh populations into 96-well ultralow-adhesion plates at 10, 100, 1000, and 5000 cells per well with 100 µl of the following medium conditions:

    • a. CSC medium.

    • b. MFCM.

      • i. Prepare MCFM before treatment as follows:

        1. Seed 7.5 × 105 18Co cells in 75-cm2 flasks and incubate overnight.

        2. The next day wash cells twice with PBS and incubate for 24 hr with 10 ml of CSC medium without EGF and FGF-basic.

        3. The next day collect MFCM and clear by centrifugation for 5 min at 1400 RPM.

        4. Use undiluted.

        5. The level of HGF present in MFCM will be determined by an ELISA following manufacturer's instructions.

  4. Incubate at 37°C for 2 hr.

    • a. After this incubation period, the plates with cells and medium will be placed in a styrofoam container and transported to the facility performing mouse injection/monitoring (∼30 min).

  5. Afterwards mix cells and medium (100 µl) with Matrigel (100 µl) at a 1:1 ratio and inject subcutaneously into the right flank of female 8–15 week old Nude mice (20–30 grams) using a sterile 25 G needle and 1 ml syringe.

    • a. Clean injection site by brief scrubbing with an isopropyl alcohol pad.

    • b. ‘Tent’ the skin of the mouse by gentle pinching with fingers and pulling upwards.

    • c. During injection, syringe is inserted into subcutaneous tissue and briefly aspirated to ensure the absence of backflow before the contents are fully injected and the needle removed from the injection site.

  6. Blindly check mice weekly for a total of 9 weeks. Record if/when tumors become detectable.

    • a. Monitor by palpitation and caliper measurement of depilated flanks.

      • i. Caliper measurements will be used to evaluate tumor volumes from their first appearance onwards.

      • ii. Calculate tumor volume as (length × width2)/2.

    • d. Confirm tumor presence at endpoint of study by necropsy.

Deliverables

  • ■ Data to be collected:

    • ○ 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-GFPlow cells compared to TOP-GFPhgh cells.

      2. TOP-GFPlow cells compared to TOP-GFPlow 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-GFPintermediate or TOP-GFPlow 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).

Sensitivity Calculations performed with G*Power software, version 3.1.7 (Faul et al., 2007).

Group 1Group 2Detectable effect size d*A priori powerGroup 1 sample sizeGroup 2 sample size
CD133 from TOP-GFPlowCD133 from TOP-GFPhigh0.05341880.0%10,00010,000
CD24 from TOP-GFPlowCD24 from TOP-GFPhigh0.05341880.0%10,00010,000
CD29 from TOP-GFPlowCD29 from TOP-GFPhigh0.05341880.0%10,00010,000
CD44 from TOP-GFPlowCD44 from TOP-GFPhigh0.05341880.0%10,00010,000
CD166 from TOP-GFPlowCD166 from TOP-GFPhigh0.05341880.0%10,00010,000
  1. *

    This is the effect size that can be detected with 80% power and with a sample size of 10,000 cells analyzed per group.

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).

  • 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).

Dataset being analyzedTotal N95% CIlowerEstimate95% CIupper
TOP-GFPlow9699.0663.3440.49
TOP-GFPhigh962.841.941.32
TOP-GFPlow + HGF968.775.803.84
TOP-GFPlow + MFCM968.975.933.92
TOP-GFPlow + HGF + PHA-66575296155.7598.1561.85
TOP-GFPlow + MFCM + PHA-66575296478266.21148.26
TOP-GFPwhole969.146.043.99
TOP-GFPwhole + PHA-665752968.265.473.62

Test family

  • ■ Chi-square test, differences between any of the groups: Bonferroni correction: alpha error = 0.01667 (corrected for the three clones from the CSC cultures).

Power calculations performed with R software, version 3.1.2 (R Development Core Team, 2014).

Groupsχ2 test statisticCohen's wA priori powerTotal sample size
TOP-GFPlow, TOP-GFPhigh, TOP-GFPlow + HGF, TOP-GFPlow + MFCM, TOP-GFPlow + HGF + PHA-665752 TOP-GFPlow + MFCM + PHA-665752, TOP-GFPwhole, TOP-GFPwhole + PHA-6657527310.97561499.9%768 (8 groups)

Test family

  • ■ Chi-square test, pairwise differences between groups: Bonferroni correction: alpha error = 0.002778.

Power calculations performed with R software, version 3.1.2 (R Development Core Team, 2014).

Group 1Group 2χ2 test statisticCohen's wA priori powerGroup 1 sample sizeGroup 2 sample size
TOP-GFPlowTOP-GFPhigh1730.94923299.9%9696
TOP-GFPlowTOP-GFPlow + HGF1140.77055299.9%9696
TOP-GFPlowTOP-GFPlow + MFCM1020.72886999.9%9696
TOP-GFPlow + HGFTOP-GFPlow + HGF + PHA-6657521530.89267999.9%9696
TOP-GFPlow + MFCMTOP-GFPlow + MFCM + PHA-6657521860.98425199.9%9696
TOP-GFPwholeTOP-GFPwhole + PHA-6657520.212*0.251143*80.0%*9696
  1. *

    A sensitivity calculation was performed since the original data showed a non-significant effect. This is the χ2 test statistic and effect size that can be detected with 80% power.

Protocol 4

Summary of original data (obtained from Figure 7E) performed with R software, version 3.1.2 (R Development Core Team, 2014).

  • 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).

  • 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.

Dataset being analyzed (C100.B5)total N95% CIlowerEstimate95% CIupper
TOP-GFPlow2418,841.86939.22555.8
TOP-GFPhigh1892.237.115.1
TOP-GFPlow + MFCM24789.3310.9122.6

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).

Groupsχ2 test statisticCohen's wA priori powerTotal sample size
TOP-GFPlow, TOP-GFPhigh, TOP-GFPlow + MFCM75.71.07096799.9%48 (3 groups)

Test family

  • ■ Chi-square test, pairwise differences between groups: Bonferroni correction: alpha error = 0.025.

Power calculations performed with R software, version 3.1.2 (R Development Core Team, 2014).

Group 1Group 2χ2 test statisticCohen's wA priori powerGroup 1 sample sizeGroup 2 sample size
TOP-GFPlowTOP-GFPhigh65.21.24594699.9%1616
TOP-GFPlowTOP-GFPlow + MFCM260.73598097.3%1616

Summary of original data (obtained from Figure 7E) performed with R software, version 3.1.2 (R Development Core Team, 2014).

Dataset being analyzed (C100.G7)total N95% CIlowerEstimate95% CIupper
TOP-GFPlow24InfinityInfinity12,238
TOP-GFPhigh182499961370
TOP-GFPlow + MFCM24523723521057

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).

Groupsχ2 test statisticCohen's wA priori powerTotal sample size
TOP-GFPlow, TOP-GFPhigh, TOP-GFPlow + MFCM26.50.63365298.19%48 (3 groups)

Test family

  • ■ Chi-square test, pairwise differences between groups: Bonferroni correction: alpha error = 0.025.

Power calculations performed with R software, version 3.1.2 (R Development Core Team, 2014).

Group 1Group 2χ2 test statisticCohen's wA priori powerGroup 1 sample sizeGroup 2 sample size
TOP-GFPlowTOP-GFPhigh21.30.71214096.3%1616
TOP-GFPlowTOP-GFPlow + MFCM17.50.60380788.0%1616

References

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10
  11. 11
  12. 12
  13. 13
  14. 14
  15. 15
  16. 16
    R: A language and environment for statistical computing. R foundation for statistical computing
    1. R Development Core Team
    (2014)
    R: A language and environment for statistical computing. R foundation for statistical computing, http://www.R-project.org/.
  17. 17
  18. 18
  19. 19
  20. 20

Decision letter

  1. Richard J Gilbertson
    Reviewing Editor; Cambridge Cancer Center, CRUK Cambridge Institute, United Kingdom

eLife posts the editorial decision letter and author response on a selection of the published articles (subject to the approval of the authors). An edited version of the letter sent to the authors after peer review is shown, indicating the substantive concerns or comments; minor concerns are not usually shown. Reviewers have the opportunity to discuss the decision before the letter is sent (see review process). Similarly, the author response typically shows only responses to the major concerns raised by the reviewers.

Thank you for sending your work entitled “Registered report: Wnt activity defines colon cancer stem cells and is regulated by the microenvironment” for consideration at eLife. Your article has been favourably evaluated by Sean Morrison (Senior Editor), Richard J Gilbertson (Reviewing Editor), and three reviewers, one of whom, Jan Paul Medema, has agreed to share his identity.

The Reviewing Editor and the reviewers discussed their comments before reaching this decision, and the Reviewing Editor has assembled the following comments that focus on the conduct of important additional experiments that will be required for further consideration of your manuscript.

Reviewer #1:

The current manuscript provides a set-up to repeat experiments performed several years ago by my team. The authors have done a careful job in listing their set-up and experimental design. Nevertheless, there are a couple of remaining concerns that need to be addressed/adjusted.

1) The name of the TOP-GFP vector is listed incorrectly. This is not a b-cat-IRES-GFP but a TOP-CMV-GFP vector.

2) In protocol 3 DMSO is added to the control. This is unlikely to have an effect, but we did not perform such an addition.

3) The titration of cells in the LD in protocol 3 is uncertain. Different spheroid cultures from individual patients have different clonogenic potential. In some case we find 1 in 4, in other cases we find 1 in 100. If the current experiments will be performed on a low clonogenic CSC (i.e. low CSC faction) the authors will need to extend their titration curves.

4) The biggest worry for the reproducibility study is the simple use of just one culture. As is clear from our study and is clear from others, there is a variability in the cultures that can be derived from colon cancers. For instance, HCT-116 and Co56 in our study do not display the same intense difference in clonogenicity and this is also reported in the study. To perform all these studies with just one culture introduces a risk into the reproducibility study. I would therefore suggest using multiple primary cultures from different patients. In addition, to facilitate the reproduction it may be wise to re-investigate the possibility of studying the clonogenic differences in one of the original cultures. This will require extensive MTAs but may facilitate the studies tremendously. I will be happy to re-invest time in checking whether this is an option. In any case, the authors should consider to use a control for clonogenic studies using AC133 staining, which has been reported by many groups to identify CSCs in spheroids, but also does not uniformaly do so (i.e. HCT116 does not work). If this fails to identify a clonogenic fraction, the TOP-GFP is also not likely to work for that particular culture.

Alternative would be to use classical cell lines to identify the top high/low difference.

5) The production of HGF by Co18 should be validated to ascertain that the TFCM is truly reproducing the original studies.

6) The MFCM stimulation of TOP-GFPlow cells prior to injection was performed in pure MFCM generated as described. So not admixed 1:2 with CSC medium.

7) The authors should cite other data using HGF/MET to study cancer stemness (i.e. J. Rich in Cancer Research and Stassi in Cell Stem Cell).

Reviewer #2:

This study aims to reproduce the key experiments from the article by Vermeulen et. al., 2010 (Nature Cell Biology). Vermeulen et al. present experiments that suggest that the cancer microenviroment regulates WNT signalling, which in turn defines colon cancer stem cells.

For the reproducibility project the authors described the replication plan of key experiments. The experiments they have chosen were originally reported in Figures 2F, 6D and 7E: I agree with the choice for Figures 6D and 7E, but I feel that two other figures should be reproduced instead of Figure 2F.

In Figure 2, Vermeulen describes the generation of TopGFP transduced clones and their subsequent analysis. The analysis presented in Figure 2 goes along the following scheme:

1) Transduction of crc spheroid cells.

2) Sorting GFPHigh and GFPLow cells (Figure 2A).

3) In-vitro clonogenic assay on GFPHigh and GFPLow cells (Fig 2A, 2B and 2D).

4) Correlation of GFPHigh with proposed cancer stem cell markers (CD133, CD24/CD29, CD44/CD166) (Figure 2F).

To me the key experiment and the basis for the study is the observation that TopGFPHigh cells show a greater clonogenic potential than TopGFPLow cells. This observation directly links WNT signaling to tumor initiating potential/cancer stem cells. Whether or not these cells express proteins that are suggested to mark cancer stem cells is not relevant as cancer stem cells are functionally defined.

By repeating Figure 2A, the in-vitro clonogenic assay, the key experiment will be reproduced. In addition to the in-vitro potential, the study should also replicate the in vivo clonogenic experiment described by Vermeulen in Figure 3A.

I suggest you repeat the following experiments:

1) In-vitro clonogenic potential (Figure 2A)

2) In-vivo clonogenic potential (Figure 3A)

3) Restore clonogenic potential of GFPLow cells with MFCM and HGF in vitro (Figure 6D)

4) Restore clonogenic potential of GFPLow cells with MFCM and HGF in vivo (Figure 7E)

(Results obtained from Figures 2A and 3A are controls for 6D and 7E)

In addition to this, the authors propose to generate a single spheroidal culture. Given the heterogeneity of colorectal cancer I suggest they make at least three independent cultures. Vermeulen also confirmed the results in different spheroid cultures. Including more patients will make this replication effort more effective and the results more sound.

Reviewer #3:

I am only commenting on the statistical part of this paper, Appendix A – Power Calculations.

1) For protocols 3 and 4, the authors proposed to compare means for multiple groups and do pairwise comparisons using Chi-square tests. There is no detail of the Chi-square test in the paper. As far as I know, the conventional tests for both scenarios are not a Chi-square test. The authors should provide the explicit test statistic they use as well as the assumptions of the test.

2) In Appendix A, under protocol 2, the variances of the two population distributions are not specified. Also, “the experiment will be performed with each of the 3 different TOP-GFP CSC cultures”. The multiple testing for the 3 scenarios is not addressed.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

Thank you for resubmitting your article entitled “Registered report: Wnt activity defines colon cancer stem cells and is regulated by the microenvironment” for further consideration at eLife. Your revised Registered Report has been evaluated by Sean Morrison (Senior Editor).

Two of the three reviewers asked you to examine at least three colon cancer lines. You agreed to include one more, for a total of two. This is helpful, but at a minimum you should use three lines.

We are also disappointed that there are no in vivo tumorigenesis experiments, as indicated by Reviewer 2 (“the study should also replicate the in vivo clonogenic experiment described by Vermeulen in Figure 3A”). In vivo tumorigenesis experiments are the gold standard for drawing conclusions regarding cancer stem cells. The markers you plan to examine are controversial and of uncertain value, so no firm conclusions will be possible irrespective of what you observe with the markers. Nonetheless, if you add a third line to make the clonogenicity experiments in culture more robust, as indicated above, we would be prepared to move forward with acceptance.

https://doi.org/10.7554/eLife.07301.002

Author response

Reviewer #1:

The current manuscript provides a set-up to repeat experiments performed several years ago by my team. The authors have done a careful job in listing their set-up and experimental design. Nevertheless, there are a couple of remaining concerns that need to be addressed/adjusted.

1) The name of the TOP-GFP vector is listed incorrectly. This is not a b-cat-IRES-GFP but a TOP-CMV-GFP vector.

We have corrected this in the revised manuscript.

2) In protocol 3 DMSO is added to the control. This is unlikely to have an effect, but we did not perform such an addition.

Thank you for this information. We have omitted the DMSO in the revised manuscript.

3) The titration of cells in the LD in protocol 3 is uncertain. Different spheroid cultures from individual patients have different clonogenic potential. In some case we find 1 in 4, in other cases we find 1 in 100. If the current experiments will be performed on a low clonogenic CSC (i.e. low CSC faction) the authors will need to extend their titration curves.

Thank you for this suggestion. We have included an additional step where each clone will be assessed to determine if the titration of cells is appropriate to detect the expected clonogenic potential. If the titration requires adjusting then this will be implemented prior to the experiment as outlined.

4) The biggest worry for the reproducibility study is the simple use of just one culture. As is clear from our study and is clear from others, there is a variability in the cultures that can be derived from colon cancers. For instance, HCT-116 and Co56 in our study do not display the same intense difference in clonogenicity and this is also reported in the study. To perform all these studies with just one culture introduces a risk into the reproducibility study. I would therefore suggest using multiple primary cultures from different patients. In addition, to facilitate the reproduction it may be wise to re-investigate the possibility of studying the clonogenic differences in one of the original cultures. This will require extensive MTAs but may facilitate the studies tremendously. I will be happy to re-invest time in checking whether this is an option. In any case, the authors should consider to use a control for clonogenic studies using AC133 staining, which has been reported by many groups to identify CSCs in spheroids, but also does not uniformaly do so (i.e. HCT116 does not work). If this fails to identify a clonogenic fraction, the TOP-GFP is also not likely to work for that particular culture.

Alternative would be to use classical cell lines to identify the top high/low difference.

Thank you for the suggestions. We are including the Co100 culture that was largely utilized in the original report as an additional group for the in vitro clonogenicity experiment. Also, each clone generated will be assessed for CD133 (AC133), CD29, CD24, and CD166 expression (Protocol 2), which will provide the additional assessment of clonogenic potential

5) The production of HGF by Co18 should be validated to ascertain that the MFCM is truly reproducing the original studies.

Thank you for this suggestion. We have included an ELISA to test for HGF levels in the MFCM that will be used.

6) The MFCM stimulation of TOP-GFPlow cells prior to injection was performed in pure MFCM generated as described. So not admixed 1:2 with CSC medium.

We have corrected this in the revised manuscript.

7) The authors should cite other data using HGF/MET to study cancer stemness (i.e. J. Rich in Cancer Research and Stassi in Cell Stem Cell).

We have expanded the Introduction to include other studies.

Reviewer #2:

[…] I suggest you repeat the following experiments:

1) In-vitro clonogenic potential (Figure 2A)

2) In-vivo clonogenic potential (Figure 3A)

3) Restore clonogenic potential of GFPLow cells with MFCM and HGF in vitro (Figure 6D)

4) Restore clonogenic potential of GFPLow cells with MFCM and HGF in vivo (Figure 7E)

(Results obtained from Figures 2A and 3A are controls for 6D and 7E)

We agree that the clonogenic potential of TOP-GFPhigh vs TOP-GFPlow cells is an important experiment, however it is depicted in figure 6D as well as Figure 2A. It includes TOP-GFPhigh and TOP-GFPlow cells without HGF or MFCM treatment, thus the control is built into the experimental design and is included as a planned comparison. This is also true of Figure 7E, which includes the TOP-GFPhigh and TOP-GFPlow cells without MFCM, which is included in both figures 6D and 7E. While this will require multiple tests to be performed, such as between TOP-GFPhigh and TOP-GFPlow and then TOP-GFPlow and TOP-GFPlow + MFCM, this has been corrected for by adjusting the alpha error.

We also agree with the point that cancer stem cells will be functionally defined by assessing the difference in clonogenic potential between TOP-GFPlow and TOP-GFPhigh cells. However, we feel the inclusion of Figure 2F allows for an additional assessment of the profile of the clones by evaluating the association with the markers originally used, which will provide another means of how this replication attempt compares to the original study.

In addition to this, the authors propose to generate a single spheroidal culture. Given the heterogeneity of colorectal cancer I suggest they make at least three independent cultures. Vermeulen also confirmed the results in different spheroid cultures. Including more patients will make this replication effort more effective and the results more sound.

Thank you for the suggestion. We are including the Co100 culture that was largely utilized in the original report as an additional group for the in vitro clonogenicity experiment. This will allow for a direct study of the clonogenic differences in one of the original cultures as well as a newly derived one.

Reviewer #3:

I am only commenting on the statistical part of this paper, Appendix A – Power Calculations.

1) For protocols 3 and 4, the authors proposed to compare means for multiple groups and do pairwise comparisons using Chi-square tests. There is no detail of the Chi-square test in the paper. As far as I know, the conventional tests for both scenarios are not a Chi-square test. The authors should provide the explicit test statistic they use as well as the assumptions of the test.

The reviewer is correct that there are no details of a Chi-square test in the original paper. There were no inferential statistics used. However, the software originally used for the determining the clonal frequency is the extreme limiting dilution analysis (ELDA), which is the limdil function of the statmod package in R. This is the same test that will be used in the analysis of the replication results. The test can also be seen on the website of authors of the analysis (http://bioinf.wehi.edu.au/software/elda/index.html). This approach uses the asymptotic Chi-square approximation to the log-ratio, which is applicable in these contexts as described in their paper (Hu and Smyth, 2009). We have updated the manuscript to make this clearer.

2) In Appendix A, under protocol 2, the variances of the two population distributions are not specified. Also,the experiment will be performed with each of the 3 different TOP-GFP CSC cultures. The multiple testing for the 3 scenarios is not addressed.

For the power calculations of protocol 2, the variances of the two population distributions are unknown, thus sensitivity calculations were performed with the anticipated number of cells to be analysed with the detectable effect size reported.

We have adjusted the alpha error for testing the three clones, which is further adjusted for the multiple comparisons within each clone.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

Two of the three reviewers asked you to examine at least three colon cancer lines. You agreed to include one more, for a total of two. This is helpful, but at a minimum you should use three lines.

We agree and have included the generation and analysis of an additional primary colon cancer line in the revised manuscript.

We are also disappointed that there are no in vivo tumorigenesis experiments, as indicated by Reviewer 2 (the study should also replicate the in vivo clonogenic experiment described by Vermeulen in Figure 3A). In vivo tumorigenesis experiments are the gold standard for drawing conclusions regarding cancer stem cells. The markers you plan to examine are controversial and of uncertain value, so no firm conclusions will be possible irrespective of what you observe with the markers. Nonetheless, if you add a third line to make the clonogenicity experiments in culture more robust, as indicated above, we would be prepared to move forward with acceptance.

We agree that the in vivo tumorigenesis experiements are valuable to include in this replication and protocol 4 is a direct replication of Figure 7E. This includes testing the in vivo clonogenic potential of the given clone (by comparing the TOP-GFPlow and TOP-GFPhigh conditions, similar to Figure 3A), while also testing the clonogenic potential of GFPlow cells with MFCM (restoring the clonogenic potential of GFPlow cells). However, we are not including these as separate experiments (thus why it is not in reference to Figure 3A), but rather performing multiple tests (and correcting for them) to test both hypotheses.

The included markers, while controversial, are being included to allow for a comparison of the replication attempt clones to the originally reported results. While it is not feasible, and beyond the scope, for this project to examine the clonogenic potential of marker positive and negative cells, we could potentially include other markers as additional exploratory analysis and welcome any additional suggestions.

https://doi.org/10.7554/eLife.07301.003

Article and author information

Author details

  1. James Evans

    PhenoVista Biosciences, San Diego, California
    Contribution
    JE, Drafting or revising the article
    Competing interests
    JE: PhenoVista Biosciences is a Science Exchange associated laboratory.
  2. Anthony Essex

    PhenoVista Biosciences, San Diego, California
    Contribution
    AE, Drafting or revising the article
    Competing interests
    AE: PhenoVista Biosciences is a Science Exchange associated laboratory.
  3. Hong Xin

    Explora BioLabs, San Diego, California
    Contribution
    HX, Drafting or revising the article
    Competing interests
    HX: Explora BioLabs is a Science Exchange associated laboratory.
  4. Nurith Amitai

    Explora BioLabs, San Diego, California
    Contribution
    NA, Drafting or revising the article
    Competing interests
    NA: Explora BioLabs is a Science Exchange associated laboratory.
  5. Lindsey Brinton

    University of Virginia, Charlottesville, Virginia
    Contribution
    LB, Drafting or revising the article
    Competing interests
    No competing interests declared.
  6. Erin Griner

    University of Virginia, Charlottesville, Virginia
    Contribution
    EG, Drafting or revising the article
    Competing interests
    No competing interests declared.
  7. Reproducibility Project: Cancer Biology

    Contribution
    RP:CB, Conception and design, Drafting or revising the article
    For correspondence
    tim@cos.io
    Competing interests
    RP:CB: EI, FT, JL: Employed by and hold shares in Science Exchange Inc.
    1. Elizabeth Iorns, Science Exchange, Palo Alto, California
    2. William Gunn, Mendeley, London, United Kingdom
    3. Fraser Tan, Science Exchange, Palo Alto, California
    4. Joelle Lomax, Science Exchange, Palo Alto, California
    5. Timothy Errington, Center for Open Science, Charlottesville, Virginia

Funding

Laura and John Arnold Foundation

  • Reproducibility Project: Cancer Biology

The Reproducibility Project: Cancer Biology is funded by the Laura and John Arnold Foundation, provided to the Center for Open Science in collaboration with Science Exchange. The funder had no role in study design or the decision to submit the work for publication.

Acknowledgements

The Reproducibility Project: Cancer Biology core team would like to thank the original authors, in particular Jan Paul Medema and Giorgio Stassi, for generously sharing critical information and reagents to ensure the fidelity and quality of this replication attempt. We thank Courtney Soderberg at the Center for Open Science for assistance with statistical analyses. We would also like to thank the following companies for generously donating reagents to the Reproducibility Project: Cancer Biology; American Type Culture Collection (ATCC), Applied Biological Materials, BioLegend, Charles River Laboratories, Corning Incorporated, DDC Medical, EMD Millipore, Harlan Laboratories, LI-COR Biosciences, Mirus Bio, Novus Biologicals, Sigma–Aldrich, and System Biosciences (SBI).

Reviewing Editor

  1. Richard J Gilbertson, Reviewing Editor, Cambridge Cancer Center, CRUK Cambridge Institute, United Kingdom

Publication history

  1. Received: March 3, 2015
  2. Accepted: August 1, 2015
  3. Version of Record published: August 19, 2015 (version 1)

Copyright

© 2015, Evans et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 1,092
    Page views
  • 183
    Downloads
  • 4
    Citations

Article citation count generated by polling the highest count across the following sources: PubMed Central, Crossref, Scopus.

Comments

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)

Further reading

  1. L Vermeulen et al.
    1
    1. Cancer Biology
    Curated by Roger Davis et al.
    Collection Updated

    Investigating reproducibility in preclinical cancer research.

    1. Developmental Biology and Stem Cells
    Jason Karch et al.
    Research Advance Updated