Bulk RNA sequencing dataset of Claudin-low breast cancer cell lines with Neuropilin-1 knockdown

Abstract

Triple-negative breast cancers (TNBC) are a particularly aggressive breast cancer subtype with poor prognosis and high relapse rates. Due to a lack of identified targeted therapies, chemotherapy currently remains as the primary treatment for TNBC. Approximately 25-39% of TNBC are claudin-low breast cancers, which are mainly defined by low expression of cell-cell adhesion proteins and enrichment of mesenchymal signatures. Functional studies have demonstrated the potential role of the transmembrane-coreceptor, Neuropilin-1 (NRP1) in regulating the progression of these tumours. However, there have been no high-throughput studies to date that comprehensively investigate NRP1-modulated cell-signalling across multiple claudin-low cell lines. Therefore, we treated HS578T, MDA-MB-231 and SUM159PT claudin-low cell lines with either a non-targeting (NT) control or two NRP1-targeting small-interfering RNA (siRNA) or short-hairpin RNA (shRNA) sequences and followed this with bulk-RNA sequencing. We present this comprehensive transcriptomic dataset which provides a valuable resource for understanding both the transcriptomic landscape of claudin-low breast cancer and NRP1-regulated signalling pathways. Therefore, paving the way for future studies of its potential as a therapeutic target.

Fig. 1

Schematic summary of dataset. Schematic summarising all cell lines and NRP1-targeting siRNA and shRNA conditions within this dataset, in addition to the number of biological replicates and experiment timepoints.

Fig. 2

RT-qPCR validation of NRP1 knockdown in claudin-low breast cancer cell lines. mRNA expression (2^−∆Ct) of (a) total NRP1 and (b) transmembrane-NRP1 in SUM159PT, HS578T and MDA-MB-231 cells following 72 hr transfection of siRNA (siNRP1#3 or #5) or lentiviral transduction of shRNA (shNRP1#3 and #5) and corresponding non-targeting (NT) controls, determined by RT-qPCR. Normalised to RPL32. N = 3. P-value determined by one-way analysis of variance (ANOVA) and Dunnett’s post-hoc multiple comparisons test. NS = non-significant, *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001. Error bars = SEM.

Fig. 3

Dataset QC validation of %GC content, STAR mapping and Kraken2. Data shown is post-trimming. (a) Histogram of average %GC content for each sample generated by FastQC v0.11.9. (b) Box plot of percentage (%) of uniquely mapped, multimapped or unmapped reads to the human reference derived from the STAR aligner output. (c) Total STAR input reads (million paired reads) per sample with individual mapping categories as derived from the STAR aligner output. (d) Counts per million (cpm, calculated with respect to the total STAR input reads) for microbial domains (Bacteria, Eukaryota, Viruses, Archaea) and (e) mycoplasma as determined by Kraken2 v2.0.9beta. All plots were generated using the R package ggplot2 (v.4.0.0).

Fig. 4. Multidimensional scaling (MDS) of all samples in the claudin-low BrCa dataset.

(a) Gene-level and (b) transcript-level MDS plot based on TMM-normalised counts. Circles, squares and triangles represent SUM159PT, MDA-MB-231 and HS578T claudin-low breast cancer cell line samples, respectively. Colouring is by treatment group, with either siNT, siNRP1#3, siNRP1#5, shNT, shNRP1#3 or shNRP1#5. Data point numbers represent independent biological replicates. MDS analysis was performed using the R package edgeR (v.4.4.2) and plotted with ggplot2 (v.4.0.0).

Fig. 5. Multidimensional scaling (MDS) of the claudin-low BrCa dataset by cell line.

Gene-level MDS plots for (a) SUM159PT, (b) MDA-MB-231, (c) HS578T as well as transcript-level MDS plots for (d) SUM159PT, (e) MDA-MB-231, (f) HS578T claudin-low cell lines, based on TMM-normalised counts, coloured by treatment group, with either siNT, siNRP1#3, siNRP1#5, shNT, shNRP1#3 or shNRP1#5. Data point numbers represent independent biological replicates. MDS analysis was performed using the R package edgeR (v.4.4.2) and plotted with ggplot2 (v.4.0.0).

Fig. 6

RNA-seq validation of NRP1 gene and transcript variant knockdown in claudin-low breast cancer cell lines. (a) Box plot of NRP1 gene expression (FPKM) and (b) bar chart of the isoform percentages (%) of NRP1 transcript variants in SUM159PT, MDA-MB-231 and HS578T cell lines treated with either siNT, siNRP1#3, siNRP1#5 or shNT, shNRP1#3 or shNRP1#5. (c) Schematic of NRP1 gene locus (as per Ensembl.v.114), showing the mapping sites of the total and transmembrane NRP1 primers, as well as the RNAi mapping sites. Note that the sequences for siNRP1#5 and shNRP1#5 are identical, whereas siNRP1#3 and shNRP1#3 sequences differ from each other, with their mapping sites being offset by about 20 nucleotides and hence, are targeting a slightly different subset of transcript variants. Plots were generated using the R packages ggplot2 (v.4.0.0) and transPlotR (v.0.0.2).