Targeted transcript quantification in single disseminated cancer cells after whole transcriptome amplification

Abstract

Gene expression analysis of rare or heterogeneous cell populations such as disseminated cancer cells (DCCs) requires a sensitive method allowing reliable analysis of single cells. Therefore, we developed and explored the feasibility of a quantitative PCR (qPCR) assay to analyze single-cell cDNA pre-amplified using a previously established whole transcriptome amplification (WTA) protocol. We carefully selected and optimized multiple steps of the protocol, e.g. re-amplification of WTA products, quantification of amplified cDNA yields and final qPCR quantification, to identify the most reliable and accurate workflow for quantitation of gene expression of the ERBB2 gene in DCCs. We found that absolute quantification outperforms relative quantification. We then validated the performance of our method on single cells of established breast cancer cell lines displaying distinct levels of HER2 protein. The different protein levels were faithfully reflected by transcript expression across the tested cell lines thereby proving the accuracy of our approach. Finally, we applied our method to breast cancer DCCs of a patient undergoing anti-HER2-directed therapy. Here, we were able to measure ERBB2 expression levels in all HER2-protein-positive DCCs. In summary, we developed a reliable single-cell qPCR assay applicable to measure distinct levels of ERBB2 in DCCs.

Fig 1. Relative qPCR quantification of single cell transcripts.

(A) Stable and uniform gene expression (depicted as the percentage of samples tested positive for selected transcripts) of twelve genes examined by endpoint PCRs of WTA products obtained from three sample sets including cell pools and single cells. The red dashed line indicates the threshold for further testing (60%). (B) Uniformity of gene expression (%) of the remaining eight candidate reference genes as measured by qPCR. (C) Stability of gene expression depicted as -ΔΔCp (log2-transformed ratios) calculated for POLR2A vs the remaining candidate genes. The expression of G6PD could not be measured in single cells (n.a. = not assessed). SC = single-cell WTA products, Pools = WTA products generated from a pool of cells. (D) HER2 expression in BT-474 and MCF-7 breast cancer cells analyzed by FACS (left panel) and microscopically (right panels). HER2 expression was 62-fold higher in BT-474 as compared to MCF-7 cells. Median FITC (BT-474) = 24,659; Median FITC (MCF-7) = 399. Microscope settings corrections: Brightness: +20%, contrast: -40%; Scale bar 20 μm. (E) Relative quantification of ERBB2 expression at the single-cell level in BT-474 and MCF-7 cells calculated separately using single reference genes as indicated. -ΔΔCp were calculated and plotted for every single-cell qPCR measurement. Mean ± SD of -ΔΔCp values; Unpaired t-test with Welch’s correction; *** p<0.001, **** p<0.0001.

Fig 2. Novel qPCR-based workflow for highly accurate gene expression analysis in single cells.

(A) Quantification of the ERBB2 gene expression levels in MCF-7 and BT-474 single cells using absolute qPCR quantification. WTA products underwent either double-stranded cDNA (dscDNA) reconstitution (left panel) or purification using the QIAquick PCR Purification Kit (right panel) prior to the normalization of cDNA input and qPCR. Cp values were converted to log10 copy numbers using an external standard curve. Mean ± SD; Unpaired t-test with Welch’s correction; **** p<0.0001. (B) Correlation of ERBB2 expression levels obtained using the relative quantification strategy for the diluted WTA samples as compared to the WTA products subjected to dscDNA reconstitution (left panel) or purification (right panel). Each point represents one -ΔΔCp value calculated for one ERBB2-reference gene pair (-ΔΔCp for all target/reference gene combinations were plotted). Pearson’s correlation coefficient R. **** p<0.0001. (C) Workflow of the established qPCR assay for profiling gene expression levels in single cells using the absolute quantification strategy.

Fig 3. Highly accurate analysis of differential expression facilitated by the new qPCR assay.

HER2 expression analysis in three breast cancer cell lines (MCF-10A, ZR-75-1 and MDA-MB-453) assessed microscopically (A) and by flow cytometry (B). (A) Brightness: +20%, contrast: -40%. Scale bar indicates 20 μm. (B) Median FITC (MCF-10A) = 436; Median FITC (ZR-75-1) = 1,529; Median FITC (MDA-MB-453) = 3,575; 3.5-fold (MCF-10A vs. ZR-75-1) and 2.3-fold (ZR-75-1 vs. MDA-MB-453) increase in HER2 expression levels. Colored lines on the x-axis indicate the signal intensities used as thresholds for sorting of cell populations prior to single-cell isolation and analysis. (C) Quantification of ERBB2 expression at the single-cell level was conducted following the newly established protocol (Fig 2C). Cp values were converted to log10 copy numbers using an external standard curve. Mean ± SD; Tukey’s multiple comparisons test was applied, n.s. = not significant, * p<0.05, ** p<0.001. (D) Correlation between measured HER2 protein and ERBB2 gene expression levels. Log10 converted median fluorescent intensity (MFI) values derived from FACS analysis and log10 converted copy numbers calculated using the absolute quantification method are plotted. Pearson’s correlation coefficient R, * p<0.05.

Fig 4. Single-cell qPCR analysis utilizing re-amplified WTA products.

(A,B) Fragment size distribution of primary WTA and corresponding WTA re-amplification products (generated with either the CP2-15C or the CP2-9C primer) of one representative BT-474 single cell assessed by Bioanalyzer assay. (A) 2–9: Re-amplified WTA products generated using the indicated primer sequence for re-amplification and dscDNA-synthesis. 4–9: Samples re-amplified with different annealing temperatures (T_A). (B) Electropherogram of selected samples. (C,D) Correlation between qPCR results conducted with primary and re-amplified WTA using CP2-15C (C) and CP2-9C (D) primers for re-amplification. (E) Absolute quantification of ERBB2 transcript levels in MCF-10A, ZR-75-1 and MDA-MB-453 cells using re-amplified single-cell WTA products (generated using the CP2-9C primer). (F) Cp values of ERBB2 expression generated for MCF-10A single cells after re-amplification using CP2-9C primer. Red, green and blue represent three different replicates.

Fig 5. HER2/ERBB2 expression in single cells derived from a clinical patient.

(A) Cells from a pleural effusion of a metastatic breast cancer patient with a HER2-positive tumor were stained using anti-EPCAM-Cy3 and anti-HER2-FITC antibodies. Cells showing a clear membranous staining pattern were picked. Scale bar indicates 20 μm. (B) ERBB2 expression was examined in primary WTA products by endpoint PCR. (C,D) Quantitative PCR analysis of ERBB2 expression was performed using re-amplified cDNA. (C) Amplification curves obtained for HER2+ and HER2- cells were labeled with red and gray colors, respectively. (D) Quantification of ERBB2 expression at the single-cell level was conducted following the newly established protocol (Fig 2C). Cp values were converted to log10 copy numbers using an external standard curve. Mean ± SD; one-way ANOVA testing with Tukey’s multiple comparisons test. Color code as in C, **** p<0.0001.