Replication Study: Wnt activity defines colon cancer stem cells and is regulated by the microenvironment

Abstract

As part of the Reproducibility Project: Cancer Biology we published a Registered Report (Evans et al., 2015), that described how we intended to replicate selected experiments from the paper 'Wnt activity defines colon cancer stem cells and is regulated by the microenvironment' (Vermeulen et al., 2010). Here, we report the results. Using three independent primary spheroidal colon cancer cultures that expressed a Wnt reporter construct we observed high Wnt activity was associated with the cell surface markers CD133, CD166, and CD29, but not CD24 and CD44, while the original study found all five markers were correlated with high Wnt activity (Figure 2F; Vermeulen et al., 2010). Clonogenicity was highest in cells with high Wnt activity and clonogenic potential of cells with low Wnt activity were increased by myofibroblast-secreted factors, including HGF. While the effects were in the same direction as the original study (Figure 6D; Vermeulen et al., 2010) whether statistical significance was reached among the different conditions varied. When tested in vivo, we did not find a difference in tumorigenicity between high and low Wnt activity, while the original study found cells with high Wnt activity were more effective in inducing tumors (Figure 7E; Vermeulen et al., 2010). Tumorigenicity, however, was increased with myofibroblast-secreted factors, which was in the same direction as the original study (Figure 7E; Vermeulen et al., 2010), but not statistically significant. Finally, we report meta-analyses for each results where possible.

Keywords: Reproducibility Project: Cancer Biology; Wnt signaling pathway; cancer biology; cancer stem cell; human; metascience; mouse; replication; reproducibility.

Figure 1.. Analysis of CSC marker expression in TOP-GFP cultures.

(A) Representative images of the three independent single-cell-cloned CSC cultures, lentivirally transduced with TOP-GFP. Phase contrast (top) and fluorescence microscopy (bottom) for each of the cultures indicated. Bar = 90 µm. (B) Single parameter histograms for GFP intensity for each of the TOP-GFP single-cell-cloned CSC cultures with the TOP-GFP^low (10% lowest) and TOP-GFP^high (10% highest) populations indicated. (C) Single parameter histograms for the indicated cell surface markers for each of the indicated cultures. Gray denotes TOP-GFP^low (10% lowest) and green denotes TOP-GFP^high (10% highest) populations. (D) Density plots for CD29/CD24 and CD44/CD166 from TOP-GFP^low (gray) and TOP-GFP^high (green) populations of each culture. Additional details for this experiment can be found at https://osf.io/tfy28/.

Figure 1—figure supplement 1.. Flow cytometry gating strategy.

Representative density plots of gating strategy to assess cell surface markers from TOP-GFP^low and TOP-GFP^high populations. Forward scatter area (FSC-A) and PerCP-Cy5.5 was used to gate on viable cells (PI negative cells), followed by forward verses side scatter area (FSC-A vs SSC-A) to identify cells of interest and exclude debris, which were then analyzed by FSC-A and forward scatter width (FSC-W), and then SSC-A and side scatter width (SSC-W) to exclude doublet cells. From the single-cell population, SSC and FITC were used to gate on the TOP-GFP^low (10% lowest) and TOP-GFP^high (10% highest) populations. TOP-GFP^low and TOP-GFP^high populations were then assessed for PE and APC to detect the fluorophores conjugated to antibodies against the cell surface markers analyzed in this study. Additional details for this experiment can be found at https://osf.io/tfy28/.

Figure 2.. Clonogenicity assay of TOP-GFP cultures.

A limiting-dilution assay was performed on the TOP-GFP^low, TOP-GFP^high, or TOP-GFP^whole populations of the three indicated TOP-GFP cultures. Cells were left untreated, or treated with 25 ng/ml HGF, 1:2 dilution of MFCM, or 500 nM PHA-665752 (PHA), as indicated. The bar graphs present the clonogenic potential of each culture with error bars representing 95% confidence intervals (y-axis is log₂ scale). This experiment was performed once for each culture. See Materials and methods and Registered Report (Evans et al., 2015) for details on limiting-dilution statistics and scheme. Planned contrast between TOP-GFP^low vs TOP-GFP^high: E450 (χ² = 39.8, uncorrected p=2.82×10⁻¹⁰, corrected p=1.69×10⁻⁹); CSC1 (χ² = 4.82, uncorrected p=0.028, corrected p=0.169); Co100 (χ² = 7.59, uncorrected p=0.0059, corrected p=0.035). Planned contrast between TOP-GFP^low vs TOP-GFP^low + HGF: E450 (χ² = 1.49, uncorrected p=0.223, corrected p>0.99); CSC1 (χ² = 0.337, uncorrected p=0.562, corrected p=0.99); Co100 (χ² = 12.7, uncorrected p=3.70×10⁻⁴, corrected p=0.0022). Planned contrast between TOP-GFP^low vs TOP-GFP^low + MFCM: E450 (χ² = 1.96, uncorrected p=0.162, corrected p=0.969); CSC1 (χ² = 4.18, uncorrected p=0.041, corrected p=0.245); Co100 (χ² = 28.7, uncorrected p=8.26×10⁻⁸, corrected p=4.96×10⁻⁷). Planned contrast between TOP-GFP^low + HGF vs TOP-GFP^low + HGF + PHA: E450 (χ² = 0.376, uncorrected p=0.540, corrected p>0.99); CSC1 (χ² = 34.0, uncorrected p=5.64×10⁻⁹, corrected p=3.39×10⁻⁸); Co100 (χ² = 5.13, uncorrected p=0.024, corrected p=0.141). Planned contrast between TOP-GFP^low + MFCM vs TOP-GFP^low + MFCM + PHA: E450 (χ² = 61.0, uncorrected p=5.71×10⁻¹⁵, corrected p=3.43×10⁻¹⁴); CSC1 (χ² = 43.5, uncorrected p=4.14×10⁻¹¹, corrected p=2.48×10⁻¹⁰); Co100 (χ² = 17.6, uncorrected p=2.67×10⁻⁵, corrected p=1.60×10⁻⁴). Planned contrast between TOP-GFP^whole vs TOP-GFP^whole + PHA: E450 (χ² = 68.3, uncorrected p=1.43×10⁻¹⁶, corrected p=8.56×10⁻¹⁶); CSC1 (χ² = 72.2, uncorrected p=1.96×10⁻¹⁷, corrected p=1.17×10⁻¹⁶); Co100 (χ² = 20.2, uncorrected p=6.91×10⁻⁶, corrected p=4.14×10⁻⁵). Additional details for this experiment can be found at https://osf.io/k9vce/.

Figure 2—figure supplement 1.. Pilot of clonogenicity assay.

A limiting-dilution assay was performed on the TOP-GFP^low, TOP-GFP^high, or TOP-GFP^whole populations of the three indicated TOP-GFP cultures. Cells were left untreated. The bar graphs present the clonogenic potential of each culture with error bars representing 95% confidence intervals (y-axis is log₂ scale). The pilot experiment was performed once for each culture. See Methods and Registered Report (Evans et al., 2015) for details on limiting-dilution scheme. Additional details for this experiment can be found at https://osf.io/k9vce/.

Figure 3.. Meta-analyses of each effect.

Effect size and 95% confidence interval are presented for Vermeulen et al. (2010), the results from this replication study (RP:CB), and a random effects meta-analysis of the effects. Cohen’s ω is a standardized measure of the association between the cells tested and clonogenic, or tumorigenic, frequency. The higher the value, the stronger the association, with an effect size of zero indicating there was no association. Sample sizes used in Vermeulen et al. (2010) and RP:CB are reported under the study name. (A) Comparison of clonogenic frequency between the indicated treated, or untreated, populations of TOP-GFP CSC cultures. TOP-GFP^low vs TOP-GFP^high (meta-analysis p=0.094); TOP-GFP^low vs TOP-GFP^low + HGF (meta-analysis p=0.110); TOP-GFP^low vs TOP-GFP^low + MFCM (meta-analysis p=0.047); TOP-GFP^low + HGF vs TOP-GFP^low + HGF + PHA (meta-analysis p=0.218); TOP-GFP^low + MFCM vs TOP-GFP^low + MFCM + PHA (meta-analysis p=0.085); TOP-GFP^whole vs TOP-GFP^whole + PHA (meta-analysis p=0.498). (B) Comparison of frequency of tumorigenicity between the indicated treated, or untreated, populations of TOP-GFP CSC cultures injected into mice. TOP-GFP^low vs TOP-GFP^high (meta-analysis p=0.330); TOP-GFP^low vs TOP-GFP^low + MFCM (meta-analysis p=0.033). Additional details for these meta-analyses can be found at https://osf.io/g4ewk/.