Supplement
Material & Methods
FACS Analysis
CD24/CD44 expression analysis:
To further identify where CD24 and CD44 correlate within the DiI+ population,
6 week DiI labeled, unsorted BxPC-3 cells were stained with mouse anti-human CD44-PE/Cy7 and CD24-FITC (both from Abcam) as previously described, and the co-expression of both FITC and PE/Cy7 was analyzed against PE (DiI) intensity.
Microarray
This experiment was performed at Norwegian Microarray Consortium in Oslo, Norway, using the Illumina Whole-Genome Gene Expression system. BxPC-3 cells were labelled with DiI and cultured for 6 weeks, as previously described. Sorted fractions (n=4/sample) were then used for the analysis, as well as the appropriate controls.
Determined intensities and ratios (including background subtraction and normalization), calculated statistical confidences (p-values) and determined differential gene expression were compiled using the Illumina BeadStudio software.
An output list of significant changes between DiI+/SCC and DiI-/FCC of selected genes associated with EMT, metastasis, CSC regulation, surface markers, proliferation/differentiation and apoptotic protection can be found in supplement table 1.
Results
DiI Intensity is Not Correlated with Increased CD24/CD44 Expression
To resolve the correlation between CD24+/CD24- and DiI+/SCC, 3 color FACS analysis was performed. As seen in figure S1,BxPC-3 DiI+/SCC were separated by gates to define intensity of DiI and thus cell cycle speed; P5 was defined as DiI+/lowSCC, P6 as DiI+/midSCC and P7 as DiI+/highSCC. Next the DiI+/SCC populations were visualized in reference to CD24 and CD44 expression, resulting again in approximately 60% CD24+/CD44+, leaving about 30% of the population CD24-/CD44+. These populations were then gated to visualize where (and if) they correlate to DiI intensity. When P8, which represents the DiI+/SCC CD24-/CD44+ population and P10, which represents the CD24+/CD44+ population were displayed on the original DiI sorting graph, no correlation was found between DiI intensity and CD24/CD44 expression, but rather an overlap of the two populations’ was found. Similarly when we preformed the CD133 and ALDH analysis, there was also no correlation found between DiI intensity and CD133 or ALDH positivity (data not shown).
Gene Expression Corroborates Functional Assays
To complement the RT-PCR data and to provide a more extensive differential analysis of the BxPC-3 DiI+/SCC and DiI-/FCC populations, expression profiling was carried out based on the Illumina whole-genome gene expression system. For data analysis, a specific focus was made on genes related to cell proliferation, apoptotic evasion, metastatic progression, EMT, cancer stem cell markers and developmental/stem cell regulation pathways. Changes in key transcription factors notably involved these processes, confirm the previous data accumulated here and further supports the role of DiI+/SCC in functioning as a CSC-like cell. An output list of significant changes between the DiI+/SCC as compared to the DiI-/FCC population are presented in table S1.
Supplementary Figure Legends
Supplementary Figure 1
(a) CD24/CD44 expression verses DiI intensity. BxPC-3 DiI+/SCC population was further grouped by DiI intensity (P5: DiI+/lowSCC; P6: DiI+/midSCC; P7: DiI+/highSCC) then analyzed against CD24 and CD44 expression. Additionally, the DiI+/SCC population was segregated to CD24-/CD44+ (P8) and CD24+/CD44+ (P10) subpopulations and the location of each group was visualized against DiI intensity. Results show no correlation of DiI intensity to CD24 expression in DiI+/SCC, but rather an overlap of the CD24-/CD44+ and CD24+/CD44+ subpopulations.
Supplement Table 1
Fold expression changes in BxPC-3 DiI+/SCC as compared to DiI-/FCC for selected genes after full microarray gene expression analysis.
Supplement Table 2
Primers and corresponding annealing temperatures used in this study.